condition stringclasses 1
value | messages listlengths 2 2 | source dict | synthetic_audit dict |
|---|---|---|---|
direct | [
{
"content": "You will be presented with a conversation involving multiple people. Your goal is to determine the kinship between two specific participants based on the dialogue. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Example: Speaker A: Do you re... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed the names (Chandler -> Maya), modified the dialogue content (divorce context shifted to a different tone), replaced the 'house-boy' reference with a general split, and updated the labels to match the parent/child relationship (2) instead of the original response's potentially ... |
direct | [
{
"content": "Analyze the provided conversation between multiple people to determine the kinship between two specific speakers. Use the following mapping for your answer: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Example: Person A: I can't believe it's been twenty years since... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed all names (Rachel -> Leo, etc.), replaced the conflict (marriage/love vs. secrets/ice cream), and modified the dialogue structure while maintaining the parent/child relationship logic (grandparent/grandchild falls under the child/parent category '2' in the context of familial ... |
direct | [
{
"content": "Analyze the provided conversations to determine the familial connection between specific individuals. Use the following mapping: '0' for spouses, '1' for siblings, and '2' for parent/child. Conversation A: Person 1: I'm so glad we finally moved into this house together. I love you so much. Person ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Removed all references to specific TV show characters (Ross, Rach), specific plot points (shell necklaces, trash chutes), and specific dialogue. Replaced with generic synthetic conversations that mirror the same relationship logic (spousal, sibling) to maintain the classification task... |
direct | [
{
"content": "You will analyze a conversation involving multiple people to determine the familial bond between two specific individuals. Use the following mapping for your answer: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/offspring. Example: Person A: Do you remember when we first m... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced the 'Friends' TV show dialogue with a synthetic scenario involving a confusing wedding memory. Changed all speaker identifiers, removed specific names (Ross, Rach), and reworded the instructions while maintaining the classification labels (0=spousal, 1=siblings, 2=parent/chil... |
direct | [
{
"content": "Part 1. Task Overview: You will analyze a conversation between multiple people to determine the familial link between two specific participants. Use the following labels: '0' for spouses/partners, '1' for siblings, and '2' for parent/child. Part 2. Sample: Person A: Do you remember where we went f... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Removed all references to the original TV show dialogue (Friends), replaced specific plot points (bachelorette/bachelor parties, strippers) with a generic anniversary gift scenario, changed all names (Joey, Ross, Mona), and rewrote the instructions and examples while maintaining the c... |
direct | [
{
"content": "Analyze the provided conversation between multiple people to determine the familial connection between two specific individuals. Use the following mapping for your answer: '0' for partners/spouses, '1' for siblings, and '2' for parent/child. Example 1 Speaker A: Welcome everyone to this celebratio... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Removed all references to 'Friends' characters (Phoebe, Mike, Ross, Joey), specific plot points (trash chutes), and distinctive dialogue. Replaced them with generic scenarios (a wedding, a domestic argument, a sibling discussion) while maintaining the logic of the classification task ... |
direct | [
{
"content": "Analyze the provided conversations between multiple people to identify the familial connection between two specific individuals. Use the following mapping for your answer: '0' for spouses/partners, '1' for siblings, and '2' for parent/child. -------- Example 1: Person A: Hey everyone, I want you t... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The original task of familial relationship identification was preserved. The complex dialogue from the source (which featured multiple characters and a specific plot about past lovers/spouses) was replaced with a new, synthetic scenario involving an art gallery owner and her husband. ... |
direct | [
{
"content": "You will be presented with a transcript of a conversation involving multiple people. Your task is to identify the familial connection between two specific individuals mentioned or participating in the text. Use the following numeric codes for your answer: '0' for spouses, '1' for siblings, and '2'... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Ross, Gloria Tribbiani, etc.) with generic labels (Person A, B, C) or new synthetic names (Marcus, Arthur, Elena). Rewrote the dialogue scenarios to remove all references to the source material (Friends TV show) while maintaining the logical structure req... |
direct | [
{
"content": "Part 1. Task Overview: You will be presented with a conversation involving multiple people. Your goal is to identify the familial connection between two specific individuals mentioned. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Part 2. ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed the dialogue entirely to remove references to 'Frank Buffay' and 'Chuck Magioni'. Replaced the 'ice cube tray' scenario with an office confrontation over a photograph. Maintained the logic where Person B is seeking their father, which preserves the 'parent/child' (label 2) rel... |
direct | [
{
"content": "Analyze the provided conversations between multiple people to determine the familial link between two specific individuals. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. [Q]: Person A: I can't believe you're still bringing this up. Person B... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Rachel, Emily, Emma, Joey, Mary-Angela, Chandler, Janice) with generic identifiers (Person A, B, Marcus). Rewrote the dialogue scenarios entirely to remove references to specific TV show tropes (Friends) and specific plot points (lip fillers, koala bears,... |
direct | [
{
"content": "Analyze the provided conversation between multiple people. Your goal is to identify the kinship between two specific participants mentioned in the text. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. [Example 1] Person A: Hey, Leo. Person B:... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all names (Phoebe, Gram, Ross, Emily, Amy) with generic identifiers or new names (Leo, Sarah). Removed specific locations (Albany, Ithaca, Burma) and brands (Ralph Lauren). Rewrote the dialogue scenarios to maintain the familial relationship logic (parent/child, sibling) whil... |
direct | [
{
"content": "Analyze the provided conversation between multiple people to determine the familial connection between two specific participants. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Conversation: Person A: Guess what? Marcus just ghosted me witho... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The original text contained several dialogue snippets from a TV show (Friends). I replaced the complex, multi-part prompt with a single, cohesive synthetic dialogue that preserves the 'sibling' relationship logic. All character names (Ross, Bob, Joey, Monica, Chandler, Rachel, Jill) w... |
direct | [
{
"content": "Analyze the provided conversation among several people to determine the familial connection between two specific individuals. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent-child. Example 1 Speaker A: Did your folks enjoy the news about the wedding... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all characters (Chandler) with generic identifiers (Speaker A, B, C, D) or new names (Marcus). Rewrote the dialogue scenarios entirely—changing catering a party to organizing a library and engagement stories to wedding news—while maintaining the logic that Speaker D and Speak... |
direct | [
{
"content": "instruction: You are provided with a transcript of a conversation between multiple people. Your task is to determine the familial relationship between two specific individuals mentioned or participating in the dialogue. Use the following mapping: '0' for spouses/partners, '1' for siblings, and '2'... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all names (Susan, Ben, Rachel, Ross, Chandler, Richard Burke) with synthetic alternatives (Maya, Leo, Sarah, Mark, Julian, Mr. Henderson). Changed specific cultural references (M*A*S*H, racquetball, twinkie) to generic activities (photo albums, hiking, general gossip). Rewrot... |
direct | [
{
"content": "TASK DEFINITION: You will be presented with a conversation involving multiple people. The dialogue contains clues about the familial connection between two specific participants. Your goal is to identify that relationship based on the following mapping: '0' for spouses, '1' for siblings, and '2' f... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed all names (Emma -> Leo, Chandler -> Julian, Ross -> Simon, Rachel -> Maya), locations (Vermont/Woodford Inn -> Oregon/Pine Ridge Lodge, The Plaza -> Grand Hotel), and specific cultural references (Maxim magazine -> travel magazine, specific wine years -> generic vintage wine).... |
direct | [
{
"content": "TASK DEFINITION: You will be provided with a transcript of a conversation involving multiple people. Your goal is to identify the familial relationship between two specific participants mentioned in the text. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Removed all references to the TV show 'Friends' (characters Joey, Monica, Phoebe, Rachel, Ross, Chandler, etc.). Changed specific locations (Nevada, China) to generic or different ones (Vegas, France, Peru). Replaced unique plot points (wedding dress, specific divorce stories, chicken... |
direct | [
{
"content": "Analyze the provided conversation between multiple people to determine the familial connection between two specific participants. Use the following mapping for your answer: '0' for spouses/partners, '1' for siblings, and '2' for parent/child. [EX Q]: Person A: I'll be right back. Person B: Did you... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Phoebe, Ross, Rachel, Meg, Joey) with generic identifiers (Person A, B, C) or new synthetic names (Julian, Clara, Sarah, Elena). Changed specific locations (Nevada, Vegas) to new ones (Italy, Atlantic City). Rewrote the dialogue to remove specific 'Friend... |
direct | [
{
"content": "Analyze the provided conversation between multiple people to determine the familial connection between two specific participants. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Example 1: Speaker A: Hello! Speaker B: Hi there. Speaker A: Do ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The task of identifying familial relationships (siblings) was preserved. All names (Monica, Silvia, Pheebs, etc.) were replaced with generic or new names (Sarah, Marcus, Jamie). The specific plot points (dollhouse, wrestling, hair comments) were replaced with new scenarios (record col... |
direct | [
{
"content": "You are provided with a conversation involving multiple people. Your task is to determine the family connection between two specific individuals mentioned. Use the following mapping: '0' for partners/spouses, '1' for siblings, and '2' for parent/child. Example: Q: Person A: Do you remember when we... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed names (Alice/Frank to Sarah), altered the dialogue scenario (Home Ec/stain to baking/cake), updated the example conversation, and maintained the class label '1' as the relationship is still between siblings.",
"attempt": 1,
"heuristic": {
"candidate_5gram_overlap_count... |
direct | [
{
"content": "Teacher: You will be provided with a conversation between multiple people. Your goal is to determine the familial connection between two specific speakers based on the context. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Teacher: Do you ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The original dialogue about divorce and a trip to London was replaced with a new scenario involving legal separation and a different emotional tone. The example was changed from a wedding anniversary to a child starting school. All specific locations and unique phrases were removed wh... |
direct | [
{
"content": "Analyze the provided conversations between multiple people to determine the familial connection between two specific individuals. Use the following mapping for your answer: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. [Q]: Person A: I can't believe we finally bough... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all dialogue content to remove original phrasing and specific references (honeymoon, shell necklace, Mike, Jack, Judy). Changed identifiers from 'Speaker' to 'Person'. Maintained the classification logic (0=spouse, 1=sibling, 2=parent/child) and the few-shot structure to pres... |
direct | [
{
"content": "You are provided with a conversation involving multiple people. Your task is to determine the familial connection between two specific speakers based on the dialogue. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Example: Q: Person A: Do y... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The original dialogue involving specific character dynamics and references to 'Chandler' and 'spray-on tans' was replaced with a generic sibling argument about childhood and clothing. The label '1' (siblings) was preserved as the synthetic dialogue explicitly mentions 'when we were ki... |
direct | [
{
"content": "You are provided with a conversation involving multiple people. Your task is to determine the familial connection between two specific speakers based on the context. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Example: Speaker A: Do you ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The scenario was rewritten from selling a house with asbestos to cleaning out a garage workshop. The kinship terms were updated (e.g., 'Dad' to 'Pops'), and the specific childhood memories (Easy Bake Ovens) were replaced with sports trophies to remove distinctive phrases while maintai... |
direct | [
{
"content": "You will be presented with a conversation involving multiple people. Your goal is to identify the familial connection between two specific individuals mentioned in the text. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Example 1: Person A:... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Removed all references to 'Friends' characters (Ross, Rachel, Monica, Chandler, Joey, Phoebe), specific locations (Nevada, China), and the original dialogue scripts. Replaced them with generic, synthetic conversation snippets that maintain the same logic for relationship classificatio... |
direct | [
{
"content": "You will be provided with a conversation involving several people. The text contains clues about the familial ties between specific participants. Your task is to identify the relationship based on the following mapping: '0' for spouses, '1' for siblings, and '2' for parent/child. Conversation 1: P... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all original dialogue, names (Ben, Mike, Mary-Angela, etc.), and locations (Colonial Williamsburg) with generic placeholders and new synthetic scenarios. Maintained the few-shot classification logic and the label mapping (0: spouse, 1: sibling, 2: parent/child) to preserve th... |
direct | [
{
"content": "You are provided with transcripts of conversations involving multiple people. Your goal is to identify the familial bond between two specific individuals mentioned. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/offspring. -------- Question: Perso... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Ross, Gloria Tribbiani, Joey) with generic labels (Person A, B, C, D) or synthetic names (Marcus, Elena Vance). Changed specific plot points (dancing lessons, wedding dress, birth of seven children) to new scenarios (piano practice, florist, career succes... |
direct | [
{
"content": "TASK DEFINITION: You are analyzing conversations to determine the familial link between two people. Use the following mapping: '0' for spouses, '1' for siblings, and '2' for parent/child. PROBLEM: Person A: Hello? Person B: Hi, I'm calling for my father, Mr. Henderson. Person A: I'm sorry, we don'... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Dorothy, Chandler, Monica) with generic labels (Person A/B) or new names (Mr. Henderson). Changed the dialogue contexts from 'sun-dried tomatoes' and 'Colonial Williamsburg' to 'vintage car restoration'. Maintained the classification logic (0=spouse, 1=si... |
direct | [
{
"content": "TASK DEFINITION: You are provided with a conversation involving multiple people. The text contains clues about the familial connection between two specific participants. Your goal is to identify that relationship using the following mapping: '0' for spouses/partners, '1' for siblings, and '2' for ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all characters, plot points, and specific dialogue from the original source (which referenced 'Friends' characters like Ross and Rachel). Maintained the few-shot prompt structure and the label mapping (0=spouse, 1=sibling, 2=parent/child). Created entirely new scenarios invol... |
direct | [
{
"content": "Analyze the provided conversations between multiple people to determine the familial connection between two specific individuals. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. -------- Example 1: Person A: Hi everyone! Person B: Hello! Pers... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all original dialogue, names (Chandler, Ursula, Phoebe, Mike), and specific references (spray tans, X-rated video stores, beepers, J. Crew) with entirely new scenarios. Maintained the few-shot structure and the label mapping (0=spouse, 1=sibling, 2=parent/child) to preserve t... |
direct | [
{
"content": "You are provided with a conversation involving multiple people. Your task is to identify the familial bond between two specific participants based on the context. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Example 1: Person A: Welcome b... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Ross, Monica, Rachel, Joey, Nana) with synthetic names (Julian, Elena, Marcus, Sarah). Changed the location (Atlantic City) to 'the mountains'. Rewrote the dialogue to maintain the same familial relationship logic (siblings) while altering the specific pl... |
direct | [
{
"content": "You are provided with transcripts of conversations involving multiple people. Your task is to identify the familial connection between two specific individuals mentioned or participating in the text. Use the following mapping for your answer: '0' for partners/spouses, '1' for siblings, and '2' for... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Mike, Ross, Paul, Rachel) with generic identifiers (Person A, Julian, Marcus, etc.) or new synthetic names (Mr. Henderson, Sarah). Rewrote all dialogue to remove specific phrases and plot points from the original source (Friends-like scenarios) while main... |
direct | [
{
"content": "instruction: You are provided with a conversation involving two or more people. The text contains clues regarding the familial bond between two specific participants. Your task is to identify that relationship using the following mapping: '0' for spouses, '1' for siblings, and '2' for parent/child... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all names (Phoe, Gram, Frank Buffay, Chandler) with synthetic ones (Leo, Nana, Mark). Changed locations (Burma, Albany, Ithaca, Saratoga, Oneonta) to generic ones (Pacific, the city, downtown). Rewrote the dialogue to remove specific plot points (tree surgeon, pharmacist, spr... |
direct | [
{
"content": "You are provided with a transcript of a conversation involving multiple people. Your goal is to determine the familial connection between two specific individuals mentioned in the text. Use the following mapping for your answer: '0' for spouses/partners, '1' for siblings, and '2' for parent/child.... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Susan, Ben, Ross, Rachel, Danny, Judy, Jack) with generic identifiers or new synthetic names (Sarah, Mark, Elena, Leo). Changed specific references like 'M*A*S*H' and 'the tube' to 'photo album from Italy' and 'the station'. Rewrote the dialogue to mainta... |
direct | [
{
"content": "Task Description: Analyze a conversation between multiple people to determine the familial connection between two specific speakers. Use the following mapping for your answer: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Example: Dialogue: Person A: Do you remember... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced the original dialogue and examples with entirely new scenarios. Changed the names (Judy, Jack) and the setting (London tube) to generic characters and a different context. Maintained the mapping logic (0=spouse, 1=sibling, 2=parent/child) and the target label '1' (siblings) a... |
direct | [
{
"content": "Analyze the provided conversation between two or more people to determine the familial connection between two specific participants. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Example: Person A: I can't believe our daughter is already st... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed the characters' names (Ross/Jill to Mark/Julian), replaced the specific workplace department (Human Resources to Finance), rewritten the dialogue to be a brother-sister dynamic instead of sister-sister while maintaining the 'sibling' label (1), and completely rephrased the ins... |
direct | [
{
"content": "Part 1. Task Overview: You will analyze a conversation among several people to determine the familial connection between two specific participants. Use the following mapping for your answer: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Part 2. Sample: Person A: Do ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced the original conversation about 'couples night' and 'swinging' with a new scenario involving a dinner party and 'date night'. Changed all speaker labels and the specific reference to Portugal to Italy. Maintained the classification logic where '0' represents a spousal/partner... |
direct | [
{
"content": "Based on the provided guidelines and examples, determine the familial connection between two people in a conversation. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Example: Person A: Do you remember when we first moved into this house tog... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced the wedding dress/mom scenario with a cake/dad scenario. Changed speaker labels from 'Speaker 1-4' to 'Person A-D'. Rewrote the example dialogue to maintain the 'spousal' logic while removing original phrasing. Preserved the label mapping (0=spouse, 1=sibling, 2=parent/child)... |
direct | [
{
"content": "You will be provided with a task description and a demonstration. Use the demonstration to process a new case. Your goal is to identify the familial connection between two people in a conversation. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parents/c... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed the scenario from a rock star's gig and odd food to a home renovation/moving context. Replaced all names (Mike) and specific phrases. Maintained the classification logic where '0' represents a spousal relationship.",
"attempt": 1,
"heuristic": {
"candidate_5gram_overla... |
direct | [
{
"content": "You are provided with a conversation involving multiple people. Your goal is to identify the familial connection between two specific speakers based on the dialogue. Use the following mapping: '0' for spouses, '1' for siblings, and '2' for parent/child. Example: Speaker A: Should we get a cake for... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed all names (Amber/Drake/Remore/Horton/Wong to Maya/Leo/Sarah/Miller), altered the professional setting from a hospital to a general office/lobby, modified the dialogue content while maintaining the 'sibling' relationship indicator, and updated the example scenario.",
"attempt... |
direct | [
{
"content": "Part 1. Task Overview: You will analyze a conversation involving multiple people to determine the familial connection between two specific individuals. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Part 2. Demonstration: Person A: Do you th... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all names (Phoebe) with generic identifiers (Person A/B). Rewrote the example dialogue to maintain the 'spousal' logic without using the original 'anniversary/passport' text. Rewrote the exercise dialogue to maintain the 'sibling' logic (mentioning a shared father) while remo... |
direct | [
{
"content": "You will be presented with a conversation involving several people. Your goal is to identify the familial connection between two specific speakers based on the context provided. Use the following mapping: '0' for spouses, '1' for siblings, and '2' for parent/child. Example: Speaker A: Do you remem... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed the scenario from a tennis/sports argument to a video gaming argument. Replaced the specific names and dialogue while maintaining the core semantic marker ('engaged' and 'partner') to justify the 'spousal' label (0). All original phrasing and specific character names (e.g., Mo... |
direct | [
{
"content": "Analyze the provided conversation among several people to determine the familial connection between two specific participants. Use the following mapping for your answer: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. Example: Speaker A: Do you remember when we first ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced the wedding/godparent scenario with a wedding/bridal shower scenario. Changed the number of speakers and the specific dialogue cues while maintaining the logic that Speaker A and Speaker D are siblings (implied through the context of family roles in the wedding party). Preser... |
direct | [
{
"content": "Analyze the provided conversation between multiple speakers to determine the familial link between two specific individuals. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Example: Speaker A: Should we book the hotel for our honeymoon now? S... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed the dialogue entirely to remove specific references to Burma, Albany, and the 'tree surgeon' plot. Replaced 'Gram' with 'Nana' and 'Phoe' with 'Leo'. The relationship being identified is now between Speaker B and their mother, preserving the 'child/parent' label (2).",
"atte... |
direct | [
{
"content": "TASK DEFINITION: You are provided with a conversation among several people. The text contains clues about the kinship between two specific individuals. Your goal is to identify their relationship based on the following mapping: '0' for spouses/partners, '1' for siblings, and '2' for parent/child. ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The task logic and label mapping (0=spouse, 1=sibling, 2=parent/child) were preserved. All specific characters, unique dialogue (the 'Mattress King' references), and the specific conversational flow were rewritten to create synthetic equivalents. New scenarios involving a home purchas... |
direct | [
{
"content": "TASK DEFINITION: Analyze a conversation between several people to determine the familial connection between two specific individuals. Use the following mapping: '0' for partners/spouses, '1' for brothers/sisters, and '2' for parent/child. PROBLEM: Person A: I will be with you in a second. Person B... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all names (Wendy, Burke, Chandler, Mike) with generic identifiers or new synthetic names (Sterling, Sarah, Leo). Changed specific locations and scenarios (on-call doctor, catering a party, honeymoon) to new contexts (intern/daughter doctor, moving furniture, wedding trip in T... |
direct | [
{
"content": "Analyze the provided conversations between multiple people to determine the familial link between two specific individuals. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. [Q]: Person A: Excuse me, why are you trailing me through the mall? Pe... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed all character names (Ursula, Phoebe, Frank Buffay, Chuck Magioni to Clara, Maya, etc.), removed specific brand references (J. Crew, Victoria's Secret), replaced specific objects (beepers, cellular phones, ice cube trays) with generic ones, and rewrote the dialogue to maintain ... |
direct | [
{
"content": "You are provided with a conversation involving multiple people. Your task is to identify the familial connection between two specific participants based on the dialogue. Use the following mapping: '0' for spouses/partners, '1' for brothers/sisters, and '2' for parent/child. Example: Person A: Do y... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Ross, Rach, Phoebe) with generic Speaker labels. Rewrote the dialogue to maintain the 'secret marriage' plot point and the resulting spousal relationship (label 0) while changing the specific phrasing and sentence structures to remove any direct copying f... |
direct | [
{
"content": "Part 1. Task Description: You will be provided with a conversation involving multiple people. Your goal is to determine the familial connection between two specific speakers based on the dialogue. Use the following mapping: '0' for spouses, '1' for siblings, and '2' for parent/child. Part 2. Illus... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The original dialogue, which was a transcript from a TV show (Friends), was completely replaced with a synthetic game-night scenario. All character names (Monica, Chandler, etc.) and specific references (TV Guide, Michael Flatley) were removed. The core task of identifying a sibling r... |
direct | [
{
"content": "Analyze the provided conversation between multiple people to determine the familial link between two specific individuals. Use the following classification keys: '0' for spouses, '1' for siblings, and '2' for parent/child. Example: Person A: Should we book the hotel for our honeymoon now? Person B... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all names (Ben, Monica, Rach -> Leo, Sarah, Julie), changed the setting (dean's office/meeting -> pharmacy), modified the dialogue flow and specific items mentioned (Diet Coke/margaritas -> markers/crayons), and rewrote the instructional framing while preserving the categoric... |
direct | [
{
"content": "Analyze the provided conversation to determine the kinship between two specific participants. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Example: Speaker A: Should we book the hotel for our honeymoon now? Speaker B: Let's wait until we f... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The original dialogue, which featured specific characters and eccentric plot points from a TV show (Friends), was completely rewritten into a generic social introduction scene. All names (Phoebe, Theodore, Bitsy, Tom, Sue Angle) and specific references (Buick LeSabre, hepatitis, pimp)... |
direct | [
{
"content": "instruction: You are provided with transcripts of conversations involving multiple people. Your task is to identify the familial connection between two specific individuals mentioned in the text. Use the following mapping: '0' for spouses/partners, '1' for siblings, and '2' for parent/child relati... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all names (Ben -> Leo, Phoe -> Maya, Frank Buffay -> Julian Thorne), specific locations (Burma, Albany, etc. -> Pacific, small town), and distinct objects (Barbi -> plastic dinosaur) with synthetic equivalents. Rewrote the dialogue to maintain the familial relationship logic ... |
direct | [
{
"content": "Analyze the provided conversation between multiple people to determine the familial connection between two specific individuals. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Conversation: Person A: Look at little Leo sleeping so soundly...... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all original names (Ben, Monica, Ross, Geller, Steven, Andrea Waltham) with synthetic names (Leo, Marcus, Sarah, James). Rewrote all dialogue to remove specific references to 'Barbie', 'wedding traditions', and 'being raised by two women' while maintaining the conversational ... |
direct | [
{
"content": "Analyze the provided conversation between multiple people to determine the familial connection between two specific participants. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Example: Speaker A: I can't believe it's been twenty years since... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed all names (Frank -> Julian, Phoebe -> Leo, Pete -> Sam, Mark -> Toby), replaced specific plot details (stilts/cigarettes -> antique clocks/snacks), rewritten dialogue to maintain the 'half-sibling' discovery plot, and corrected the target relationship question from the origina... |
direct | [
{
"content": "Analyze the provided conversation between multiple people to determine the familial connection between two specific participants. Use the following mapping for your answer: '0' for spouses/partners, '1' for siblings, and '2' for parent/child. Example 1 Input: Person A: Look at you! You've grown so... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The original task of identifying familial relationships was preserved. All original dialogue (including the specific sitcom-style scenarios and character names like Ross, Joey, and Mona) was replaced with entirely new, generic scenarios: a parent/child interaction, a couple separating... |
direct | [
{
"content": "Analyze the provided conversations between multiple people to determine the kinship between two specific participants. Use the following mapping for your answer: '0' for spouses/partners, '1' for brothers/sisters, and '2' for parent/child. -------- Question: Person A: Listen, Sarah, I messed up...... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Amy, Ross, Rach, Phoebe) with generic names (Sarah, Mark, Julie, Chloe) or labels (Person A, B, etc.). Changed specific brand references (Ralph Lauren, Easy Bake Ovens) to generic items (gym membership, puzzles). Rewrote the dialogue to maintain the same ... |
direct | [
{
"content": "You will be presented with conversations involving multiple people. Your task is to determine the familial connection between two specific individuals mentioned or participating in the dialogue. Use the following mapping: '0' for spouses/partners, '1' for siblings, and '2' for parent/child. ------... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all characters (Ross, Rachel, Joey, Mike) with generic identifiers (Person A, B, C, D) or new names (Julian, Sarah, David). Changed locations (Atlantic City) to generic ones (mountains). Altered specific plot points (saltwater taffy, roller-skating party) to new scenarios (ca... |
direct | [
{
"content": "You are provided with a transcript of a conversation involving several people. Your task is to identify the familial relationship between two specific participants mentioned at the end of the text. Use the following mapping: '0' for spouses/partners, '1' for siblings, and '2' for parent/child. Q: ... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The task of identifying familial relationships was preserved. All character names (Geller, Rachel, Emily, Emma, Dina, Joe, Bobby Corso) were replaced with synthetic names (Thorne, Maya, Marcus, Leo, Julian). The scenarios were rewritten: the court scene was simplified, the sibling riv... |
direct | [
{
"content": "Analyze the provided conversation between several people to determine the specific familial connection between two designated speakers. Use the following mapping: '0' for partners/spouses, '1' for siblings, and '2' for parent/child. Example: Q: Person A: Should we book the cruise for our honeymoon... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Changed the business context (wine/cheese to watches/coins), replaced all names (Don, Chandler to Marcus, Leo), changed the destination (France/Loire Valley to Italy/Tuscany), and completely rewrote the dialogue while maintaining the core relationship dynamics (Speaker B and C are spo... |
direct | [
{
"content": "Analyze the provided conversation between several people to determine the familial link between two specific participants. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Example 1: Person A: The project is finished. Person B: Great, let's ce... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Removed all references to the TV show 'Friends' (characters Ross, Rachel, Phoebe; specific plot points about the Buick, hepatitis, etc.). Replaced names with generic identifiers (Person A, B, C) and new synthetic names (Elena, Mark, Sarah, Arthur, Martha). Rewrote the dialogue to main... |
direct | [
{
"content": "TASK DEFINITION: You are provided with a transcript of a conversation involving two or more people. The dialogue contains clues regarding the familial connection between two specific participants. Your goal is to identify that relationship based on the provided mapping: '0' for spouses, '1' for si... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Emma, Chandler, Frank Buffay, Chuck Magioni, Phoebe, Theodore, Bitsy, Tom, Sue Angle) with synthetic names (Leo, Marcus, Arthur Vance, Julian Thorne, Clara, Harold, Martha, Greg, Linda). Changed locations (Vermont, Woodford Inn) to (Georgia, Blue Ridge Mo... |
direct | [
{
"content": "TASK DEFINITION: You are provided with a transcript of a conversation involving multiple people. The text contains clues about the family ties between specific participants. Your goal is to identify the relationship between two designated speakers. Use the following mapping: '0' for spouses, '1' f... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "Replaced all character names (Ross, Rachel, Paul, Francis, Lilly, Ursula, Phoebe, Monica) with new synthetic names (Marcus, Mr. Sterling, Julian, Sarah, Maya, Elena, Clara). Changed specific locations (Soho, country house) to generic ones (city, lakeside cabin). Rewrote all dialogue t... |
direct | [
{
"content": "Part 1. Task Description: You will analyze a conversation between two or more people to determine the kinship between two specific individuals. Use the following mapping for your answer: '0' for spouses, '1' for siblings, and '2' for parent/child. Part 2. Sample Case: Person A: I can't believe it'... | {
"campaign": "sapient-excluded-synthetic-anonymous",
"group": "high40",
"source_path": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relationships.parquet",
"source_row_id": "sapient_cleaned/data_clustered/flan/niv2_fsopt_data__task908_dialogre_identify_familial_relat... | {
"anonymous_generation_notes": "The familial relationship identification task was preserved. The wedding anniversary example was replaced with a sibling-based example, and the father/son dialogue was replaced with a mother/child dialogue. All specific phrasing and identifiers were rewritten to be synthetic.",
"att... |
End of preview. Expand in Data Studio
sapient-synth-flan-niv2-fsopt-data-task908-dialogre-identify-familial-relationships
Chat-template-ready synthetic anonymous replacement examples for one Sapient source excluded from the DFM5 data mix.
Contents
- Format: gzip-compressed JSON Lines under
data/train.jsonl.gz - Schema:
{"messages": [{"role": "user", "content": "..."}, {"role": "assistant", "content": "..."}]} - Files:
1 - Rows:
163 - Task: synthetic anonymous instruction replacement
Generation
Rows were generated with google/gemma-4-31B-it and accepted only after a google/gemma-4-31B-it judge checked task preservation, PII absence, low textual overlap, and usefulness.
Overlap precautions: prompts asked the model to create new synthetic examples that preserve the task type and skill, not to rewrite or copy the provenance row. Candidates were rejected when the judge found substantial wording overlap, unresolved PII, or failure to preserve the task. Only accepted rows are included.
License
Released as Apache-2.0. The datasets that inspired these recreations may have different licensing conditions. These examples are fully synthetic recreations with no or minimal wording overlap and were judged to be free of PII before inclusion.
Recreate
python recreate_dataset.py --validate
python recreate_dataset.py --from-synth /path/to/synth/flan_niv2_fsopt_data_task908_dialogre_identify_familial_relationships.parquet --output /path/to/output-folder
- Downloads last month
- 25