reneeice commited on
Commit
b914a23
·
verified ·
1 Parent(s): acd2efe

Upload folder using huggingface_hub

Browse files
README.md CHANGED
@@ -25,7 +25,7 @@ import torch
25
  from transformers import AutoTokenizer, AutoModel
26
  from peft import PeftModel
27
 
28
- base = "Qwen/Qwen3-0.6B-Base"
29
  tok = AutoTokenizer.from_pretrained("reneeice/ood-editguard-qwen3-0.6b")
30
  backbone = PeftModel.from_pretrained(AutoModel.from_pretrained(base, torch_dtype=torch.bfloat16),
31
  "reneeice/ood-editguard-qwen3-0.6b")
@@ -41,9 +41,9 @@ Validation on `pangram/editlens_iclr` (held-out):
41
 
42
  | Metric | Value |
43
  |---|---|
44
- | **AUROC** (AI vs human) | **0.941** |
45
- | AUPR | 0.969 |
46
- | correlation with edit-magnitude | +0.661 |
47
 
48
  A random detector scores AUROC 0.5.
49
 
@@ -67,7 +67,7 @@ The journey, start to finish:
67
  people lightly edit their own drafts with AI. [EditLens](https://arxiv.org/abs/2510.03154)
68
  (Thai et al., 2025) reframes detection as a **continuous "extent of AI editing"**
69
  score in [0,1], and the community
70
- [`editlens-qwen3-*-repro`](https://huggingface.co/reneeice/editlens-qwen3-4b-repro)
71
  models bring it to a modern **Qwen3** backbone.
72
 
73
  3. **Apply the OOD idea to the edit-detection setting.** The insight of this work:
@@ -78,7 +78,7 @@ The journey, start to finish:
78
  | Model | What it is | Use it when |
79
  |---|---|---|
80
  | [`ood-editguard-qwen3-0.6b`](https://huggingface.co/reneeice/ood-editguard-qwen3-0.6b) ← **you are here** | **Standalone OOD AI-edit detector** — a Qwen3 backbone fine-tuned (QLoRA) with an out-of-distribution head; outputs a continuous "how AI-edited" score. | You want one self-contained model that scores text end-to-end. |
81
- | [`editlens-ood-adapter-qwen3-0.6b`](https://huggingface.co/reneeice/editlens-ood-adapter-qwen3-0.6b) | **Tiny OOD adapter** (a few MB) that snaps onto a frozen [EditLens-Qwen3](https://huggingface.co/reneeice/editlens-qwen3-4b-repro) checkpoint to add an anomaly / human-likeness score — no backbone training. | You already run EditLens and want to add an OOD score cheaply. |
82
  | [`editlens-ood-selective-guard-qwen3`](https://huggingface.co/reneeice/editlens-ood-selective-guard-qwen3) | **Reliability guard** for selective prediction — an OOD gate that abstains on inputs unlike the training distribution so the edit-score isn't trusted blindly. | You need calibrated, low-false-positive decisions and can abstain on hard cases. |
83
 
84
  > **Why three?** They trade off cost and integration: **A** is a standalone model,
@@ -100,7 +100,7 @@ baked into this family.
100
 
101
  ## How it was trained
102
 
103
- - **Backbone:** `Qwen/Qwen3-0.6B-Base`, bf16 + LoRA (rank 8, all attn+MLP projections).
104
  - **Head:** a small LayerNorm+Linear projection trained in full, with a DeepSVDD
105
  one-class objective: pull **human** embeddings toward a center `c`, push AI
106
  embeddings away. Score = oriented squared distance to `c`.
 
25
  from transformers import AutoTokenizer, AutoModel
26
  from peft import PeftModel
27
 
28
+ base = "Qwen/Qwen3-1.7B-Base"
29
  tok = AutoTokenizer.from_pretrained("reneeice/ood-editguard-qwen3-0.6b")
30
  backbone = PeftModel.from_pretrained(AutoModel.from_pretrained(base, torch_dtype=torch.bfloat16),
31
  "reneeice/ood-editguard-qwen3-0.6b")
 
41
 
42
  | Metric | Value |
43
  |---|---|
44
+ | **AUROC** (AI vs human) | **0.955** |
45
+ | AUPR | 0.977 |
46
+ | correlation with edit-magnitude | +0.723 |
47
 
48
  A random detector scores AUROC 0.5.
49
 
 
67
  people lightly edit their own drafts with AI. [EditLens](https://arxiv.org/abs/2510.03154)
68
  (Thai et al., 2025) reframes detection as a **continuous "extent of AI editing"**
69
  score in [0,1], and the community
70
+ community `editlens-qwen3-*-repro` models (search HF: `editlens qwen3 repro`)
71
  models bring it to a modern **Qwen3** backbone.
72
 
73
  3. **Apply the OOD idea to the edit-detection setting.** The insight of this work:
 
78
  | Model | What it is | Use it when |
79
  |---|---|---|
80
  | [`ood-editguard-qwen3-0.6b`](https://huggingface.co/reneeice/ood-editguard-qwen3-0.6b) ← **you are here** | **Standalone OOD AI-edit detector** — a Qwen3 backbone fine-tuned (QLoRA) with an out-of-distribution head; outputs a continuous "how AI-edited" score. | You want one self-contained model that scores text end-to-end. |
81
+ | [`editlens-ood-adapter-qwen3-0.6b`](https://huggingface.co/reneeice/editlens-ood-adapter-qwen3-0.6b) | **Tiny OOD adapter** (a few MB) that snaps onto a frozen EditLens-Qwen3 (search HF: `editlens qwen3 repro`) checkpoint to add an anomaly / human-likeness score — no backbone training. | You already run EditLens and want to add an OOD score cheaply. |
82
  | [`editlens-ood-selective-guard-qwen3`](https://huggingface.co/reneeice/editlens-ood-selective-guard-qwen3) | **Reliability guard** for selective prediction — an OOD gate that abstains on inputs unlike the training distribution so the edit-score isn't trusted blindly. | You need calibrated, low-false-positive decisions and can abstain on hard cases. |
83
 
84
  > **Why three?** They trade off cost and integration: **A** is a standalone model,
 
100
 
101
  ## How it was trained
102
 
103
+ - **Backbone:** `Qwen/Qwen3-1.7B-Base`, bf16 + LoRA (rank 8, all attn+MLP projections).
104
  - **Head:** a small LayerNorm+Linear projection trained in full, with a DeepSVDD
105
  one-class objective: pull **human** embeddings toward a center `c`, push AI
106
  embeddings away. Score = oriented squared distance to `c`.
adapter_config.json CHANGED
@@ -3,7 +3,7 @@
3
  "alpha_pattern": {},
4
  "arrow_config": null,
5
  "auto_mapping": null,
6
- "base_model_name_or_path": "Qwen/Qwen3-0.6B-Base",
7
  "bias": "none",
8
  "corda_config": null,
9
  "ensure_weight_tying": false,
@@ -30,13 +30,13 @@
30
  "rank_pattern": {},
31
  "revision": null,
32
  "target_modules": [
33
- "k_proj",
34
- "down_proj",
35
- "o_proj",
36
- "gate_proj",
37
- "q_proj",
38
  "v_proj",
39
- "up_proj"
 
 
 
 
 
40
  ],
41
  "target_parameters": null,
42
  "task_type": "FEATURE_EXTRACTION",
 
3
  "alpha_pattern": {},
4
  "arrow_config": null,
5
  "auto_mapping": null,
6
+ "base_model_name_or_path": "Qwen/Qwen3-1.7B-Base",
7
  "bias": "none",
8
  "corda_config": null,
9
  "ensure_weight_tying": false,
 
30
  "rank_pattern": {},
31
  "revision": null,
32
  "target_modules": [
 
 
 
 
 
33
  "v_proj",
34
+ "up_proj",
35
+ "q_proj",
36
+ "gate_proj",
37
+ "o_proj",
38
+ "k_proj",
39
+ "down_proj"
40
  ],
41
  "target_parameters": null,
42
  "task_type": "FEATURE_EXTRACTION",
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b5d466a9e18c99aabb8707bcff47d0b11bbad404176565bdfa82151c989b80cc
3
- size 20234120
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a164c563b092e1ae53f294f5f72404acd5a7b951559783c598d7432ef768ec3
3
+ size 34914368
added_tokens.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<|box_end|>": 151649,
9
+ "<|box_start|>": 151648,
10
+ "<|endoftext|>": 151643,
11
+ "<|file_sep|>": 151664,
12
+ "<|fim_middle|>": 151660,
13
+ "<|fim_pad|>": 151662,
14
+ "<|fim_prefix|>": 151659,
15
+ "<|fim_suffix|>": 151661,
16
+ "<|im_end|>": 151645,
17
+ "<|im_start|>": 151644,
18
+ "<|image_pad|>": 151655,
19
+ "<|object_ref_end|>": 151647,
20
+ "<|object_ref_start|>": 151646,
21
+ "<|quad_end|>": 151651,
22
+ "<|quad_start|>": 151650,
23
+ "<|repo_name|>": 151663,
24
+ "<|video_pad|>": 151656,
25
+ "<|vision_end|>": 151653,
26
+ "<|vision_pad|>": 151654,
27
+ "<|vision_start|>": 151652
28
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
ood_head.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7d4ff0553b14f0d5a78937c79bda6ac26f51f7cc6489945ee66df6f7078419dc
3
- size 1059992
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b0bc46ea5d3130a0fe3fe604e01d34ab8a1b3589bfcd2521c4769971bd6265b2
3
+ size 2116760
special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|endoftext|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
tokenizer.json CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7029094cd70eca33e2f5d6837051bd1b63789ebde3c05bcce93b0fb31c094a85
3
- size 11422928
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:574de68a0f63f2004784a421c7d42c2b2786c05cb38542d2ed3525757a1f7fde
3
+ size 11422932
tokenizer_config.json CHANGED
@@ -1,11 +1,217 @@
1
  {
 
2
  "add_prefix_space": false,
3
- "backend": "tokenizers",
4
- "bos_token": null,
5
- "clean_up_tokenization_spaces": false,
6
- "eos_token": "<|endoftext|>",
7
- "errors": "replace",
8
- "extra_special_tokens": [
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9
  "<|im_start|>",
10
  "<|im_end|>",
11
  "<|object_ref_start|>",
@@ -20,8 +226,11 @@
20
  "<|image_pad|>",
21
  "<|video_pad|>"
22
  ],
23
- "is_local": false,
24
- "local_files_only": false,
 
 
 
25
  "model_max_length": 131072,
26
  "pad_token": "<|endoftext|>",
27
  "split_special_tokens": false,
 
1
  {
2
+ "add_bos_token": false,
3
  "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ },
181
+ "151665": {
182
+ "content": "<tool_response>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": false
188
+ },
189
+ "151666": {
190
+ "content": "</tool_response>",
191
+ "lstrip": false,
192
+ "normalized": false,
193
+ "rstrip": false,
194
+ "single_word": false,
195
+ "special": false
196
+ },
197
+ "151667": {
198
+ "content": "<think>",
199
+ "lstrip": false,
200
+ "normalized": false,
201
+ "rstrip": false,
202
+ "single_word": false,
203
+ "special": false
204
+ },
205
+ "151668": {
206
+ "content": "</think>",
207
+ "lstrip": false,
208
+ "normalized": false,
209
+ "rstrip": false,
210
+ "single_word": false,
211
+ "special": false
212
+ }
213
+ },
214
+ "additional_special_tokens": [
215
  "<|im_start|>",
216
  "<|im_end|>",
217
  "<|object_ref_start|>",
 
226
  "<|image_pad|>",
227
  "<|video_pad|>"
228
  ],
229
+ "bos_token": null,
230
+ "clean_up_tokenization_spaces": false,
231
+ "eos_token": "<|endoftext|>",
232
+ "errors": "replace",
233
+ "extra_special_tokens": {},
234
  "model_max_length": 131072,
235
  "pad_token": "<|endoftext|>",
236
  "split_special_tokens": false,
vocab.json ADDED
The diff for this file is too large to render. See raw diff