README.md

metadata

license: llama2
tags:
  - mergekit
  - merge
model-index:
  - name: WinterGoddess-1.4x-70b-32k
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: AI2 Reasoning Challenge (25-Shot)
          type: ai2_arc
          config: ARC-Challenge
          split: test
          args:
            num_few_shot: 25
        metrics:
          - type: acc_norm
            value: 71.16
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ChuckMcSneed/WinterGoddess-1.4x-70b-32k
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: HellaSwag (10-Shot)
          type: hellaswag
          split: validation
          args:
            num_few_shot: 10
        metrics:
          - type: acc_norm
            value: 89.12
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ChuckMcSneed/WinterGoddess-1.4x-70b-32k
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU (5-Shot)
          type: cais/mmlu
          config: all
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 66.42
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ChuckMcSneed/WinterGoddess-1.4x-70b-32k
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: TruthfulQA (0-shot)
          type: truthful_qa
          config: multiple_choice
          split: validation
          args:
            num_few_shot: 0
        metrics:
          - type: mc2
            value: 63.87
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ChuckMcSneed/WinterGoddess-1.4x-70b-32k
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: Winogrande (5-shot)
          type: winogrande
          config: winogrande_xl
          split: validation
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 82.56
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ChuckMcSneed/WinterGoddess-1.4x-70b-32k
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GSM8k (5-shot)
          type: gsm8k
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 43.29
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ChuckMcSneed/WinterGoddess-1.4x-70b-32k
          name: Open LLM Leaderboard

This is a 32k version of Sao10K/WinterGoddess-1.4x-70B-L2, extended using method discussed here.

Quants

Thanks for GGUF, @Nexesenex!

• GGUF

Benchmarks

NeoEvalPlusN_benchmark

My meme benchmark.

Test name	WinterGoddess	WinterGoddess-32k
B	2	2.5
C	1.5	2
D	3	0
S	2.75	1.5
P	5.5	2.25
Total	14.75	8.25

Open LLM Leaderboard Evaluation Results

Leaderboard on Huggingface

Model	Average	ARC	HellaSwag	MMLU	TruthfulQA	Winogrande	GSM8K
Sao10K/WinterGoddess-1.4x-70B-L2	73.23	72.78	90.11	71.12	65.76	85	54.59
ChuckMcSneed/WinterGoddess-1.4x-70b-32k	69.4	71.16	89.12	66.42	63.87	82.56	43.29
Difference	3.83	1.62	0.99	4.7	1.89	2.44	11.3

Here the losses seem far less brutal than on my bench. It seems that extending with LongLORA kills MMLU and GSM8K performance.

Detailed results can be found here

Metric	Value
Avg.	69.40
AI2 Reasoning Challenge (25-Shot)	71.16
HellaSwag (10-Shot)	89.12
MMLU (5-Shot)	66.42
TruthfulQA (0-shot)	63.87
Winogrande (5-shot)	82.56
GSM8k (5-shot)	43.29