Safetensors
obalcells commited on
Commit
1e3cedf
·
verified ·
1 Parent(s): 8497c07

Upload LoRA model and probe head

Browse files
.gitattributes CHANGED
@@ -86,3 +86,7 @@ mistral_small_24b_linear/llama3_1_8b_validation_roc_curves.png filter=lfs diff=l
86
  mistral_small_24b_linear/mistral_small_24b_longfact_augmented_test_roc_curves.png filter=lfs diff=lfs merge=lfs -text
87
  mistral_small_24b_linear/mistral_small_24b_longfact_test_roc_curves.png filter=lfs diff=lfs merge=lfs -text
88
  mistral_small_24b_linear/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
86
  mistral_small_24b_linear/mistral_small_24b_longfact_augmented_test_roc_curves.png filter=lfs diff=lfs merge=lfs -text
87
  mistral_small_24b_linear/mistral_small_24b_longfact_test_roc_curves.png filter=lfs diff=lfs merge=lfs -text
88
  mistral_small_24b_linear/tokenizer.json filter=lfs diff=lfs merge=lfs -text
89
+ qwen2_5_7b_lora_lambda_kl_0_05/llama3_1_8b_trivia_qa_test_roc_curves.png filter=lfs diff=lfs merge=lfs -text
90
+ qwen2_5_7b_lora_lambda_kl_0_05/qwen_2_5_7b_longfact_augmented_test_roc_curves.png filter=lfs diff=lfs merge=lfs -text
91
+ qwen2_5_7b_lora_lambda_kl_0_05/qwen_2_5_7b_longfact_test_roc_curves.png filter=lfs diff=lfs merge=lfs -text
92
+ qwen2_5_7b_lora_lambda_kl_0_05/tokenizer.json filter=lfs diff=lfs merge=lfs -text
qwen2_5_7b_lora_lambda_kl_0_05/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: Qwen/Qwen2.5-7B-Instruct
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.14.0
qwen2_5_7b_lora_lambda_kl_0_05/adapter_config.json ADDED
@@ -0,0 +1,222 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": {
4
+ "base_model_class": "Qwen2ForCausalLM",
5
+ "parent_library": "transformers.models.qwen2.modeling_qwen2"
6
+ },
7
+ "base_model_name_or_path": "Qwen/Qwen2.5-7B-Instruct",
8
+ "bias": "none",
9
+ "eva_config": null,
10
+ "exclude_modules": null,
11
+ "fan_in_fan_out": false,
12
+ "inference_mode": true,
13
+ "init_lora_weights": true,
14
+ "layer_replication": null,
15
+ "layers_pattern": null,
16
+ "layers_to_transform": null,
17
+ "loftq_config": {},
18
+ "lora_alpha": 32,
19
+ "lora_bias": false,
20
+ "lora_dropout": 0.05,
21
+ "megatron_config": null,
22
+ "megatron_core": "megatron.core",
23
+ "modules_to_save": null,
24
+ "peft_type": "LORA",
25
+ "r": 16,
26
+ "rank_pattern": {},
27
+ "revision": null,
28
+ "target_modules": [
29
+ "model.layers.20.mlp.down_proj",
30
+ "model.layers.10.mlp.up_proj",
31
+ "model.layers.20.self_attn.o_proj",
32
+ "model.layers.10.mlp.down_proj",
33
+ "model.layers.3.self_attn.o_proj",
34
+ "model.layers.9.mlp.down_proj",
35
+ "model.layers.12.mlp.down_proj",
36
+ "model.layers.17.self_attn.k_proj",
37
+ "model.layers.4.mlp.up_proj",
38
+ "model.layers.17.mlp.gate_proj",
39
+ "model.layers.6.self_attn.o_proj",
40
+ "model.layers.8.self_attn.q_proj",
41
+ "model.layers.16.self_attn.k_proj",
42
+ "model.layers.17.self_attn.v_proj",
43
+ "model.layers.10.mlp.gate_proj",
44
+ "model.layers.26.mlp.up_proj",
45
+ "model.layers.10.self_attn.v_proj",
46
+ "model.layers.2.self_attn.o_proj",
47
+ "model.layers.9.mlp.up_proj",
48
+ "model.layers.2.self_attn.q_proj",
49
+ "model.layers.5.mlp.down_proj",
50
+ "model.layers.26.self_attn.v_proj",
51
+ "model.layers.25.self_attn.k_proj",
52
+ "model.layers.3.self_attn.k_proj",
53
+ "model.layers.2.self_attn.k_proj",
54
+ "model.layers.15.mlp.down_proj",
55
+ "model.layers.26.self_attn.k_proj",
56
+ "model.layers.1.self_attn.k_proj",
57
+ "model.layers.16.mlp.up_proj",
58
+ "model.layers.22.mlp.down_proj",
59
+ "model.layers.7.self_attn.q_proj",
60
+ "model.layers.18.self_attn.v_proj",
61
+ "model.layers.15.self_attn.k_proj",
62
+ "model.layers.9.mlp.gate_proj",
63
+ "model.layers.23.mlp.down_proj",
64
+ "model.layers.14.mlp.down_proj",
65
+ "model.layers.22.self_attn.v_proj",
66
+ "model.layers.4.self_attn.q_proj",
67
+ "model.layers.25.mlp.down_proj",
68
+ "model.layers.6.self_attn.q_proj",
69
+ "model.layers.6.self_attn.k_proj",
70
+ "model.layers.8.mlp.gate_proj",
71
+ "model.layers.9.self_attn.o_proj",
72
+ "model.layers.4.mlp.gate_proj",
73
+ "model.layers.18.mlp.up_proj",
74
+ "model.layers.2.mlp.down_proj",
75
+ "model.layers.1.self_attn.o_proj",
76
+ "model.layers.12.self_attn.q_proj",
77
+ "model.layers.2.mlp.up_proj",
78
+ "model.layers.17.self_attn.q_proj",
79
+ "model.layers.22.self_attn.o_proj",
80
+ "model.layers.18.self_attn.k_proj",
81
+ "model.layers.7.self_attn.o_proj",
82
+ "model.layers.23.mlp.up_proj",
83
+ "model.layers.13.self_attn.o_proj",
84
+ "model.layers.6.mlp.gate_proj",
85
+ "model.layers.13.self_attn.q_proj",
86
+ "model.layers.5.self_attn.k_proj",
87
+ "model.layers.15.self_attn.o_proj",
88
+ "model.layers.17.self_attn.o_proj",
89
+ "model.layers.8.self_attn.v_proj",
90
+ "model.layers.5.mlp.up_proj",
91
+ "model.layers.21.self_attn.k_proj",
92
+ "model.layers.11.self_attn.v_proj",
93
+ "model.layers.2.mlp.gate_proj",
94
+ "model.layers.15.self_attn.v_proj",
95
+ "model.layers.21.mlp.down_proj",
96
+ "model.layers.26.mlp.gate_proj",
97
+ "model.layers.25.mlp.gate_proj",
98
+ "model.layers.1.mlp.gate_proj",
99
+ "model.layers.3.mlp.down_proj",
100
+ "model.layers.22.self_attn.q_proj",
101
+ "model.layers.10.self_attn.o_proj",
102
+ "model.layers.3.self_attn.q_proj",
103
+ "model.layers.16.self_attn.o_proj",
104
+ "model.layers.20.self_attn.v_proj",
105
+ "model.layers.3.self_attn.v_proj",
106
+ "model.layers.18.self_attn.q_proj",
107
+ "model.layers.16.mlp.gate_proj",
108
+ "model.layers.4.self_attn.v_proj",
109
+ "model.layers.12.self_attn.o_proj",
110
+ "model.layers.17.mlp.up_proj",
111
+ "model.layers.12.mlp.up_proj",
112
+ "model.layers.4.self_attn.o_proj",
113
+ "model.layers.21.self_attn.q_proj",
114
+ "model.layers.23.mlp.gate_proj",
115
+ "model.layers.4.self_attn.k_proj",
116
+ "model.layers.15.mlp.up_proj",
117
+ "model.layers.1.self_attn.v_proj",
118
+ "model.layers.6.mlp.up_proj",
119
+ "model.layers.16.mlp.down_proj",
120
+ "model.layers.13.mlp.down_proj",
121
+ "model.layers.18.mlp.gate_proj",
122
+ "model.layers.24.mlp.gate_proj",
123
+ "model.layers.19.self_attn.o_proj",
124
+ "model.layers.7.self_attn.v_proj",
125
+ "model.layers.6.mlp.down_proj",
126
+ "model.layers.16.self_attn.q_proj",
127
+ "model.layers.25.self_attn.q_proj",
128
+ "model.layers.23.self_attn.q_proj",
129
+ "model.layers.20.self_attn.q_proj",
130
+ "model.layers.26.self_attn.q_proj",
131
+ "model.layers.25.self_attn.o_proj",
132
+ "model.layers.18.self_attn.o_proj",
133
+ "model.layers.0.self_attn.v_proj",
134
+ "model.layers.22.self_attn.k_proj",
135
+ "model.layers.7.mlp.up_proj",
136
+ "model.layers.19.self_attn.k_proj",
137
+ "model.layers.0.mlp.gate_proj",
138
+ "model.layers.21.self_attn.o_proj",
139
+ "model.layers.1.self_attn.q_proj",
140
+ "model.layers.13.self_attn.k_proj",
141
+ "model.layers.6.self_attn.v_proj",
142
+ "model.layers.3.mlp.up_proj",
143
+ "model.layers.13.mlp.gate_proj",
144
+ "model.layers.15.self_attn.q_proj",
145
+ "model.layers.22.mlp.gate_proj",
146
+ "model.layers.21.self_attn.v_proj",
147
+ "model.layers.21.mlp.up_proj",
148
+ "model.layers.20.mlp.up_proj",
149
+ "model.layers.7.self_attn.k_proj",
150
+ "model.layers.22.mlp.up_proj",
151
+ "model.layers.23.self_attn.v_proj",
152
+ "model.layers.24.mlp.down_proj",
153
+ "model.layers.24.self_attn.v_proj",
154
+ "model.layers.13.mlp.up_proj",
155
+ "model.layers.19.mlp.up_proj",
156
+ "model.layers.7.mlp.gate_proj",
157
+ "model.layers.0.self_attn.o_proj",
158
+ "model.layers.20.self_attn.k_proj",
159
+ "model.layers.24.mlp.up_proj",
160
+ "model.layers.0.mlp.down_proj",
161
+ "model.layers.8.self_attn.k_proj",
162
+ "model.layers.5.mlp.gate_proj",
163
+ "model.layers.23.self_attn.k_proj",
164
+ "model.layers.2.self_attn.v_proj",
165
+ "model.layers.12.mlp.gate_proj",
166
+ "model.layers.17.mlp.down_proj",
167
+ "model.layers.5.self_attn.v_proj",
168
+ "model.layers.25.self_attn.v_proj",
169
+ "model.layers.16.self_attn.v_proj",
170
+ "model.layers.5.self_attn.q_proj",
171
+ "model.layers.14.self_attn.k_proj",
172
+ "model.layers.19.mlp.down_proj",
173
+ "model.layers.15.mlp.gate_proj",
174
+ "model.layers.14.mlp.up_proj",
175
+ "model.layers.19.mlp.gate_proj",
176
+ "model.layers.21.mlp.gate_proj",
177
+ "model.layers.0.self_attn.q_proj",
178
+ "model.layers.12.self_attn.k_proj",
179
+ "model.layers.14.self_attn.q_proj",
180
+ "model.layers.18.mlp.down_proj",
181
+ "model.layers.5.self_attn.o_proj",
182
+ "model.layers.11.mlp.gate_proj",
183
+ "model.layers.14.self_attn.o_proj",
184
+ "model.layers.1.mlp.up_proj",
185
+ "model.layers.19.self_attn.v_proj",
186
+ "model.layers.9.self_attn.v_proj",
187
+ "model.layers.10.self_attn.k_proj",
188
+ "model.layers.24.self_attn.k_proj",
189
+ "model.layers.10.self_attn.q_proj",
190
+ "model.layers.11.mlp.up_proj",
191
+ "model.layers.8.mlp.up_proj",
192
+ "model.layers.0.mlp.up_proj",
193
+ "model.layers.24.self_attn.o_proj",
194
+ "model.layers.11.self_attn.q_proj",
195
+ "model.layers.9.self_attn.q_proj",
196
+ "model.layers.26.self_attn.o_proj",
197
+ "model.layers.8.mlp.down_proj",
198
+ "model.layers.11.self_attn.o_proj",
199
+ "model.layers.11.self_attn.k_proj",
200
+ "model.layers.0.self_attn.k_proj",
201
+ "model.layers.20.mlp.gate_proj",
202
+ "model.layers.26.mlp.down_proj",
203
+ "model.layers.7.mlp.down_proj",
204
+ "model.layers.12.self_attn.v_proj",
205
+ "model.layers.14.self_attn.v_proj",
206
+ "model.layers.19.self_attn.q_proj",
207
+ "model.layers.8.self_attn.o_proj",
208
+ "model.layers.1.mlp.down_proj",
209
+ "model.layers.3.mlp.gate_proj",
210
+ "model.layers.9.self_attn.k_proj",
211
+ "model.layers.11.mlp.down_proj",
212
+ "model.layers.23.self_attn.o_proj",
213
+ "model.layers.25.mlp.up_proj",
214
+ "model.layers.13.self_attn.v_proj",
215
+ "model.layers.4.mlp.down_proj",
216
+ "model.layers.24.self_attn.q_proj",
217
+ "model.layers.14.mlp.gate_proj"
218
+ ],
219
+ "task_type": null,
220
+ "use_dora": false,
221
+ "use_rslora": false
222
+ }
qwen2_5_7b_lora_lambda_kl_0_05/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a60796c8d5ac06a6f3ee18442154d8dd43fca202df91763b6519e4f4a96dbe27
3
+ size 155764128
qwen2_5_7b_lora_lambda_kl_0_05/added_tokens.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|box_end|>": 151649,
5
+ "<|box_start|>": 151648,
6
+ "<|endoftext|>": 151643,
7
+ "<|file_sep|>": 151664,
8
+ "<|fim_middle|>": 151660,
9
+ "<|fim_pad|>": 151662,
10
+ "<|fim_prefix|>": 151659,
11
+ "<|fim_suffix|>": 151661,
12
+ "<|im_end|>": 151645,
13
+ "<|im_start|>": 151644,
14
+ "<|image_pad|>": 151655,
15
+ "<|object_ref_end|>": 151647,
16
+ "<|object_ref_start|>": 151646,
17
+ "<|quad_end|>": 151651,
18
+ "<|quad_start|>": 151650,
19
+ "<|repo_name|>": 151663,
20
+ "<|video_pad|>": 151656,
21
+ "<|vision_end|>": 151653,
22
+ "<|vision_pad|>": 151654,
23
+ "<|vision_start|>": 151652
24
+ }
qwen2_5_7b_lora_lambda_kl_0_05/eval_metrics.jsonl ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {"qwen_2_5_7b_longfact_test/lm_loss": 0.3498629196631981, "qwen_2_5_7b_longfact_test/probe_loss": 0.944612454062868, "qwen_2_5_7b_longfact_test/sparsity": 0.09246595673127059, "qwen_2_5_7b_longfact_test/probe_threshold": 0.5, "qwen_2_5_7b_longfact_test/all_tokens_accuracy": 0.9391203196301536, "qwen_2_5_7b_longfact_test/all_tokens_precision": 0.35071555131525145, "qwen_2_5_7b_longfact_test/all_tokens_recall": 0.33921273959239634, "qwen_2_5_7b_longfact_test/all_tokens_f1": 0.3448682552872115, "qwen_2_5_7b_longfact_test/all_tokens_auc": 0.8749180157651084, "qwen_2_5_7b_longfact_test/all_tokens_optimal_threshold": 0.9415640782828283, "qwen_2_5_7b_longfact_test/all_tokens_threshold_optimized_accuracy": 0.9528612777235319, "qwen_2_5_7b_longfact_test/all_tokens_recall_at_0.1_fpr": 0.6260644889134963, "qwen_2_5_7b_longfact_test/all_tokens_true_positive_count": 37929, "qwen_2_5_7b_longfact_test/all_tokens_true_negative_count": 764999, "qwen_2_5_7b_longfact_test/all_tokens_pred_positive_count": 36685, "qwen_2_5_7b_longfact_test/all_tokens_pred_negative_count": 766243, "qwen_2_5_7b_longfact_test/all_tokens_total_samples": 802928, "qwen_2_5_7b_longfact_test/entity_tokens_accuracy": 0.6994360881753034, "qwen_2_5_7b_longfact_test/entity_tokens_precision": 0.8376847451005924, "qwen_2_5_7b_longfact_test/entity_tokens_recall": 0.33921273959239634, "qwen_2_5_7b_longfact_test/entity_tokens_f1": 0.48288545263473953, "qwen_2_5_7b_longfact_test/entity_tokens_auc": 0.8200465888374614, "qwen_2_5_7b_longfact_test/entity_tokens_optimal_threshold": 0.16702178030303028, "qwen_2_5_7b_longfact_test/entity_tokens_threshold_optimized_accuracy": 0.7524350737884622, "qwen_2_5_7b_longfact_test/entity_tokens_recall_at_0.1_fpr": 0.5104273774684278, "qwen_2_5_7b_longfact_test/entity_tokens_true_positive_count": 37929, "qwen_2_5_7b_longfact_test/entity_tokens_true_negative_count": 53752, "qwen_2_5_7b_longfact_test/entity_tokens_pred_positive_count": 15359, "qwen_2_5_7b_longfact_test/entity_tokens_pred_negative_count": 76322, "qwen_2_5_7b_longfact_test/entity_tokens_total_samples": 91681, "qwen_2_5_7b_longfact_test/entity_span_max_accuracy": 0.7995004856389621, "qwen_2_5_7b_longfact_test/entity_span_max_precision": 0.7547085201793722, "qwen_2_5_7b_longfact_test/entity_span_max_recall": 0.6520728399845022, "qwen_2_5_7b_longfact_test/entity_span_max_f1": 0.6996466431095406, "qwen_2_5_7b_longfact_test/entity_span_max_auc": 0.8738263111336242, "qwen_2_5_7b_longfact_test/entity_span_max_optimal_threshold": 0.42876025883838387, "qwen_2_5_7b_longfact_test/entity_span_max_threshold_optimized_accuracy": 0.80158179547662, "qwen_2_5_7b_longfact_test/entity_span_max_recall_at_0.1_fpr": 0.604804339403332, "qwen_2_5_7b_longfact_test/entity_span_max_true_positive_count": 5162, "qwen_2_5_7b_longfact_test/entity_span_max_true_negative_count": 9252, "qwen_2_5_7b_longfact_test/entity_span_max_pred_positive_count": 4460, "qwen_2_5_7b_longfact_test/entity_span_max_pred_negative_count": 9954, "qwen_2_5_7b_longfact_test/entity_span_max_total_samples": 14414, "epoch": 1.0, "global_step": 241, "training_progress": 1.0, "dataset_id": "qwen_2_5_7b_longfact_test"}
2
+ {"qwen_2_5_7b_longfact_augmented_test/lm_loss": 0.32481150685728966, "qwen_2_5_7b_longfact_augmented_test/probe_loss": 1.147995874615435, "qwen_2_5_7b_longfact_augmented_test/sparsity": 0.11234531300662642, "qwen_2_5_7b_longfact_augmented_test/probe_threshold": 0.5, "qwen_2_5_7b_longfact_augmented_test/all_tokens_accuracy": 0.9210031726846956, "qwen_2_5_7b_longfact_augmented_test/all_tokens_precision": 0.3536936360729133, "qwen_2_5_7b_longfact_augmented_test/all_tokens_recall": 0.39221845078271445, "qwen_2_5_7b_longfact_augmented_test/all_tokens_f1": 0.3719611799750168, "qwen_2_5_7b_longfact_augmented_test/all_tokens_auc": 0.8674433076262933, "qwen_2_5_7b_longfact_augmented_test/all_tokens_optimal_threshold": 1.0, "qwen_2_5_7b_longfact_augmented_test/all_tokens_threshold_optimized_accuracy": 0.940355038525457, "qwen_2_5_7b_longfact_augmented_test/all_tokens_recall_at_0.1_fpr": 0.5928111859769999, "qwen_2_5_7b_longfact_augmented_test/all_tokens_true_positive_count": 39478, "qwen_2_5_7b_longfact_augmented_test/all_tokens_true_negative_count": 622422, "qwen_2_5_7b_longfact_augmented_test/all_tokens_pred_positive_count": 43778, "qwen_2_5_7b_longfact_augmented_test/all_tokens_pred_negative_count": 618122, "qwen_2_5_7b_longfact_augmented_test/all_tokens_total_samples": 661900, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_accuracy": 0.6830000245984306, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_precision": 0.8968952734012975, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_recall": 0.39221845078271445, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_f1": 0.5457685664939551, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_auc": 0.8364956916506707, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_optimal_threshold": 0.14692826704545453, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_threshold_optimized_accuracy": 0.7554547019900131, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_recall_at_0.1_fpr": 0.5543593900400223, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_true_positive_count": 39478, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_true_negative_count": 41828, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_pred_positive_count": 17264, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_pred_negative_count": 64042, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_total_samples": 81306, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_accuracy": 0.7992657273268505, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_precision": 0.784524449220849, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_recall": 0.7051436851002173, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_f1": 0.7427190639704947, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_auc": 0.8818148801162915, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_optimal_threshold": 0.42960858585858597, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_threshold_optimized_accuracy": 0.8006548918436198, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_recall_at_0.1_fpr": 0.6314899782661193, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_true_positive_count": 4141, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_true_negative_count": 5937, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_pred_positive_count": 3722, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_pred_negative_count": 6356, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_total_samples": 10078, "epoch": 1.0, "global_step": 241, "training_progress": 1.0, "dataset_id": "qwen_2_5_7b_longfact_augmented_test"}
3
+ {"llama3_1_8b_trivia_qa_test/lm_loss": NaN, "llama3_1_8b_trivia_qa_test/probe_loss": 0.9877010567864943, "llama3_1_8b_trivia_qa_test/sparsity": 0.19224896117420875, "llama3_1_8b_trivia_qa_test/probe_threshold": 0.5, "llama3_1_8b_trivia_qa_test/all_tokens_accuracy": 0.8708254051571419, "llama3_1_8b_trivia_qa_test/all_tokens_precision": 0.17068645640074212, "llama3_1_8b_trivia_qa_test/all_tokens_recall": 0.36687116564417177, "llama3_1_8b_trivia_qa_test/all_tokens_f1": 0.23297944871919743, "llama3_1_8b_trivia_qa_test/all_tokens_auc": 0.7602149036036673, "llama3_1_8b_trivia_qa_test/all_tokens_optimal_threshold": 1.0, "llama3_1_8b_trivia_qa_test/all_tokens_threshold_optimized_accuracy": 0.9465094153926907, "llama3_1_8b_trivia_qa_test/all_tokens_recall_at_0.1_fpr": 0.36441717791411044, "llama3_1_8b_trivia_qa_test/all_tokens_true_positive_count": 3260, "llama3_1_8b_trivia_qa_test/all_tokens_true_negative_count": 57704, "llama3_1_8b_trivia_qa_test/all_tokens_pred_positive_count": 7007, "llama3_1_8b_trivia_qa_test/all_tokens_pred_negative_count": 53957, "llama3_1_8b_trivia_qa_test/all_tokens_total_samples": 60964, "llama3_1_8b_trivia_qa_test/entity_tokens_accuracy": 0.6302420729628367, "llama3_1_8b_trivia_qa_test/entity_tokens_precision": 0.9192928516525749, "llama3_1_8b_trivia_qa_test/entity_tokens_recall": 0.36687116564417177, "llama3_1_8b_trivia_qa_test/entity_tokens_f1": 0.524446393334795, "llama3_1_8b_trivia_qa_test/entity_tokens_auc": 0.8037175889523468, "llama3_1_8b_trivia_qa_test/entity_tokens_optimal_threshold": 0.11838600852272729, "llama3_1_8b_trivia_qa_test/entity_tokens_threshold_optimized_accuracy": 0.7337197408796454, "llama3_1_8b_trivia_qa_test/entity_tokens_recall_at_0.1_fpr": 0.5104294478527608, "llama3_1_8b_trivia_qa_test/entity_tokens_true_positive_count": 3260, "llama3_1_8b_trivia_qa_test/entity_tokens_true_negative_count": 2606, "llama3_1_8b_trivia_qa_test/entity_tokens_pred_positive_count": 1301, "llama3_1_8b_trivia_qa_test/entity_tokens_pred_negative_count": 4565, "llama3_1_8b_trivia_qa_test/entity_tokens_total_samples": 5866, "llama3_1_8b_trivia_qa_test/entity_span_max_accuracy": 0.8203321590337191, "llama3_1_8b_trivia_qa_test/entity_span_max_precision": 0.8865853658536585, "llama3_1_8b_trivia_qa_test/entity_span_max_recall": 0.7336024217961655, "llama3_1_8b_trivia_qa_test/entity_span_max_f1": 0.8028713418001104, "llama3_1_8b_trivia_qa_test/entity_span_max_auc": 0.9113396066607499, "llama3_1_8b_trivia_qa_test/entity_span_max_optimal_threshold": 0.3879616477272727, "llama3_1_8b_trivia_qa_test/entity_span_max_threshold_optimized_accuracy": 0.8424760946149975, "llama3_1_8b_trivia_qa_test/entity_span_max_recall_at_0.1_fpr": 0.7467204843592331, "llama3_1_8b_trivia_qa_test/entity_span_max_true_positive_count": 991, "llama3_1_8b_trivia_qa_test/entity_span_max_true_negative_count": 996, "llama3_1_8b_trivia_qa_test/entity_span_max_pred_positive_count": 820, "llama3_1_8b_trivia_qa_test/entity_span_max_pred_negative_count": 1167, "llama3_1_8b_trivia_qa_test/entity_span_max_total_samples": 1987, "epoch": 1.0, "global_step": 241, "training_progress": 1.0, "dataset_id": "llama3_1_8b_trivia_qa_test"}
4
+ {"qwen_2_5_7b_longfact_test/lm_loss": 0.3403156998186459, "qwen_2_5_7b_longfact_test/probe_loss": 0.9048037807860596, "qwen_2_5_7b_longfact_test/sparsity": 0.0762635828693386, "qwen_2_5_7b_longfact_test/probe_threshold": 0.5, "qwen_2_5_7b_longfact_test/all_tokens_accuracy": 0.9440310961879521, "qwen_2_5_7b_longfact_test/all_tokens_precision": 0.3892365061307041, "qwen_2_5_7b_longfact_test/all_tokens_recall": 0.324738326873896, "qwen_2_5_7b_longfact_test/all_tokens_f1": 0.3540741379558162, "qwen_2_5_7b_longfact_test/all_tokens_auc": 0.8909883076961829, "qwen_2_5_7b_longfact_test/all_tokens_optimal_threshold": 0.8556265782828287, "qwen_2_5_7b_longfact_test/all_tokens_threshold_optimized_accuracy": 0.9531788653528087, "qwen_2_5_7b_longfact_test/all_tokens_recall_at_0.1_fpr": 0.6661130006063961, "qwen_2_5_7b_longfact_test/all_tokens_true_positive_count": 37929, "qwen_2_5_7b_longfact_test/all_tokens_true_negative_count": 764999, "qwen_2_5_7b_longfact_test/all_tokens_pred_positive_count": 31644, "qwen_2_5_7b_longfact_test/all_tokens_pred_negative_count": 771284, "qwen_2_5_7b_longfact_test/all_tokens_total_samples": 802928, "qwen_2_5_7b_longfact_test/entity_tokens_accuracy": 0.6977454434397531, "qwen_2_5_7b_longfact_test/entity_tokens_precision": 0.8543978912319645, "qwen_2_5_7b_longfact_test/entity_tokens_recall": 0.324738326873896, "qwen_2_5_7b_longfact_test/entity_tokens_f1": 0.47060846308147863, "qwen_2_5_7b_longfact_test/entity_tokens_auc": 0.8382365663877721, "qwen_2_5_7b_longfact_test/entity_tokens_optimal_threshold": 0.13589015151515155, "qwen_2_5_7b_longfact_test/entity_tokens_threshold_optimized_accuracy": 0.7659166021313031, "qwen_2_5_7b_longfact_test/entity_tokens_recall_at_0.1_fpr": 0.5349468744232645, "qwen_2_5_7b_longfact_test/entity_tokens_true_positive_count": 37929, "qwen_2_5_7b_longfact_test/entity_tokens_true_negative_count": 53752, "qwen_2_5_7b_longfact_test/entity_tokens_pred_positive_count": 14416, "qwen_2_5_7b_longfact_test/entity_tokens_pred_negative_count": 77265, "qwen_2_5_7b_longfact_test/entity_tokens_total_samples": 91681, "qwen_2_5_7b_longfact_test/entity_span_max_accuracy": 0.8037324823088664, "qwen_2_5_7b_longfact_test/entity_span_max_precision": 0.7878114976560573, "qwen_2_5_7b_longfact_test/entity_span_max_recall": 0.6185586981790003, "qwen_2_5_7b_longfact_test/entity_span_max_f1": 0.6930005425935974, "qwen_2_5_7b_longfact_test/entity_span_max_auc": 0.8837167661414778, "qwen_2_5_7b_longfact_test/entity_span_max_optimal_threshold": 0.437204071969697, "qwen_2_5_7b_longfact_test/entity_span_max_threshold_optimized_accuracy": 0.8098376578326627, "qwen_2_5_7b_longfact_test/entity_span_max_recall_at_0.1_fpr": 0.6375435877566834, "qwen_2_5_7b_longfact_test/entity_span_max_true_positive_count": 5162, "qwen_2_5_7b_longfact_test/entity_span_max_true_negative_count": 9252, "qwen_2_5_7b_longfact_test/entity_span_max_pred_positive_count": 4053, "qwen_2_5_7b_longfact_test/entity_span_max_pred_negative_count": 10361, "qwen_2_5_7b_longfact_test/entity_span_max_total_samples": 14414, "epoch": 1.0, "global_step": 964, "training_progress": 1.0, "dataset_id": "qwen_2_5_7b_longfact_test"}
5
+ {"qwen_2_5_7b_longfact_augmented_test/lm_loss": 0.31472883214076014, "qwen_2_5_7b_longfact_augmented_test/probe_loss": 0.9969771905236889, "qwen_2_5_7b_longfact_augmented_test/sparsity": 0.09702653702202062, "qwen_2_5_7b_longfact_augmented_test/probe_threshold": 0.5, "qwen_2_5_7b_longfact_augmented_test/all_tokens_accuracy": 0.9277186886236591, "qwen_2_5_7b_longfact_augmented_test/all_tokens_precision": 0.3952882858072754, "qwen_2_5_7b_longfact_augmented_test/all_tokens_recall": 0.3999442727595116, "qwen_2_5_7b_longfact_augmented_test/all_tokens_f1": 0.39760264917339244, "qwen_2_5_7b_longfact_augmented_test/all_tokens_auc": 0.8851051986364417, "qwen_2_5_7b_longfact_augmented_test/all_tokens_optimal_threshold": 0.8589409722222228, "qwen_2_5_7b_longfact_augmented_test/all_tokens_threshold_optimized_accuracy": 0.9409397189907841, "qwen_2_5_7b_longfact_augmented_test/all_tokens_recall_at_0.1_fpr": 0.6396727291149501, "qwen_2_5_7b_longfact_augmented_test/all_tokens_true_positive_count": 39478, "qwen_2_5_7b_longfact_augmented_test/all_tokens_true_negative_count": 622422, "qwen_2_5_7b_longfact_augmented_test/all_tokens_pred_positive_count": 39943, "qwen_2_5_7b_longfact_augmented_test/all_tokens_pred_negative_count": 621957, "qwen_2_5_7b_longfact_augmented_test/all_tokens_total_samples": 661900, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_accuracy": 0.6896416008658648, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_precision": 0.9108688127379716, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_recall": 0.3999442727595116, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_f1": 0.5558332746602831, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_auc": 0.8556854617116961, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_optimal_threshold": 0.12335759943181823, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_threshold_optimized_accuracy": 0.7739527218163481, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_recall_at_0.1_fpr": 0.5913420132732154, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_true_positive_count": 39478, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_true_negative_count": 41828, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_pred_positive_count": 17334, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_pred_negative_count": 63972, "qwen_2_5_7b_longfact_augmented_test/entity_tokens_total_samples": 81306, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_accuracy": 0.8000595356221473, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_precision": 0.8051090700344432, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_recall": 0.6773726153103116, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_f1": 0.7357377049180328, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_auc": 0.8888790523144552, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_optimal_threshold": 0.4113794191919192, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_threshold_optimized_accuracy": 0.8064100019845207, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_recall_at_0.1_fpr": 0.644771794252596, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_true_positive_count": 4141, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_true_negative_count": 5937, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_pred_positive_count": 3484, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_pred_negative_count": 6594, "qwen_2_5_7b_longfact_augmented_test/entity_span_max_total_samples": 10078, "epoch": 1.0, "global_step": 964, "training_progress": 1.0, "dataset_id": "qwen_2_5_7b_longfact_augmented_test"}
6
+ {"llama3_1_8b_trivia_qa_test/lm_loss": 1.2490080514870792, "llama3_1_8b_trivia_qa_test/probe_loss": 0.888429161903274, "llama3_1_8b_trivia_qa_test/sparsity": 0.19553201752473634, "llama3_1_8b_trivia_qa_test/probe_threshold": 0.5, "llama3_1_8b_trivia_qa_test/all_tokens_accuracy": 0.8640181090479627, "llama3_1_8b_trivia_qa_test/all_tokens_precision": 0.19581519109820997, "llama3_1_8b_trivia_qa_test/all_tokens_recall": 0.49662576687116566, "llama3_1_8b_trivia_qa_test/all_tokens_f1": 0.28088133240804997, "llama3_1_8b_trivia_qa_test/all_tokens_auc": 0.8182136579829025, "llama3_1_8b_trivia_qa_test/all_tokens_optimal_threshold": 0.9733664772727275, "llama3_1_8b_trivia_qa_test/all_tokens_threshold_optimized_accuracy": 0.9466734466242372, "llama3_1_8b_trivia_qa_test/all_tokens_recall_at_0.1_fpr": 0.4634969325153374, "llama3_1_8b_trivia_qa_test/all_tokens_true_positive_count": 3260, "llama3_1_8b_trivia_qa_test/all_tokens_true_negative_count": 57704, "llama3_1_8b_trivia_qa_test/all_tokens_pred_positive_count": 8268, "llama3_1_8b_trivia_qa_test/all_tokens_pred_negative_count": 52696, "llama3_1_8b_trivia_qa_test/all_tokens_total_samples": 60964, "llama3_1_8b_trivia_qa_test/entity_tokens_accuracy": 0.6924650528469144, "llama3_1_8b_trivia_qa_test/entity_tokens_precision": 0.9085297418630752, "llama3_1_8b_trivia_qa_test/entity_tokens_recall": 0.49662576687116566, "llama3_1_8b_trivia_qa_test/entity_tokens_f1": 0.6422054740182467, "llama3_1_8b_trivia_qa_test/entity_tokens_auc": 0.8600766753457099, "llama3_1_8b_trivia_qa_test/entity_tokens_optimal_threshold": 0.12021583017676771, "llama3_1_8b_trivia_qa_test/entity_tokens_threshold_optimized_accuracy": 0.7877599727241732, "llama3_1_8b_trivia_qa_test/entity_tokens_recall_at_0.1_fpr": 0.6058282208588958, "llama3_1_8b_trivia_qa_test/entity_tokens_true_positive_count": 3260, "llama3_1_8b_trivia_qa_test/entity_tokens_true_negative_count": 2606, "llama3_1_8b_trivia_qa_test/entity_tokens_pred_positive_count": 1782, "llama3_1_8b_trivia_qa_test/entity_tokens_pred_negative_count": 4084, "llama3_1_8b_trivia_qa_test/entity_tokens_total_samples": 5866, "llama3_1_8b_trivia_qa_test/entity_span_max_accuracy": 0.8570709612481128, "llama3_1_8b_trivia_qa_test/entity_span_max_precision": 0.8701570680628272, "llama3_1_8b_trivia_qa_test/entity_span_max_recall": 0.8385469223007064, "llama3_1_8b_trivia_qa_test/entity_span_max_f1": 0.8540596094552929, "llama3_1_8b_trivia_qa_test/entity_span_max_auc": 0.9361077002257263, "llama3_1_8b_trivia_qa_test/entity_span_max_optimal_threshold": 0.5555555555555558, "llama3_1_8b_trivia_qa_test/entity_span_max_threshold_optimized_accuracy": 0.8580775037745345, "llama3_1_8b_trivia_qa_test/entity_span_max_recall_at_0.1_fpr": 0.8173562058526741, "llama3_1_8b_trivia_qa_test/entity_span_max_true_positive_count": 991, "llama3_1_8b_trivia_qa_test/entity_span_max_true_negative_count": 996, "llama3_1_8b_trivia_qa_test/entity_span_max_pred_positive_count": 955, "llama3_1_8b_trivia_qa_test/entity_span_max_pred_negative_count": 1032, "llama3_1_8b_trivia_qa_test/entity_span_max_total_samples": 1987, "epoch": 1.0, "global_step": 964, "training_progress": 1.0, "dataset_id": "llama3_1_8b_trivia_qa_test"}
qwen2_5_7b_lora_lambda_kl_0_05/eval_metrics_qwen2_5_7b_lora_lambda_kl=0.5.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "qwen_2_5_7b_longfact_test/lm_loss": 0.3403156998186459,
3
+ "qwen_2_5_7b_longfact_test/probe_loss": 0.9048037807860596,
4
+ "qwen_2_5_7b_longfact_test/sparsity": 0.0762635828693386,
5
+ "qwen_2_5_7b_longfact_test/probe_threshold": 0.5,
6
+ "qwen_2_5_7b_longfact_test/all_tokens_accuracy": 0.9440310961879521,
7
+ "qwen_2_5_7b_longfact_test/all_tokens_precision": 0.3892365061307041,
8
+ "qwen_2_5_7b_longfact_test/all_tokens_recall": 0.324738326873896,
9
+ "qwen_2_5_7b_longfact_test/all_tokens_f1": 0.3540741379558162,
10
+ "qwen_2_5_7b_longfact_test/all_tokens_auc": 0.8909883076961829,
11
+ "qwen_2_5_7b_longfact_test/all_tokens_optimal_threshold": 0.8556265782828287,
12
+ "qwen_2_5_7b_longfact_test/all_tokens_threshold_optimized_accuracy": 0.9531788653528087,
13
+ "qwen_2_5_7b_longfact_test/all_tokens_recall_at_0.1_fpr": 0.6661130006063961,
14
+ "qwen_2_5_7b_longfact_test/all_tokens_true_positive_count": 37929,
15
+ "qwen_2_5_7b_longfact_test/all_tokens_true_negative_count": 764999,
16
+ "qwen_2_5_7b_longfact_test/all_tokens_pred_positive_count": 31644,
17
+ "qwen_2_5_7b_longfact_test/all_tokens_pred_negative_count": 771284,
18
+ "qwen_2_5_7b_longfact_test/all_tokens_total_samples": 802928,
19
+ "qwen_2_5_7b_longfact_test/entity_tokens_accuracy": 0.6977454434397531,
20
+ "qwen_2_5_7b_longfact_test/entity_tokens_precision": 0.8543978912319645,
21
+ "qwen_2_5_7b_longfact_test/entity_tokens_recall": 0.324738326873896,
22
+ "qwen_2_5_7b_longfact_test/entity_tokens_f1": 0.47060846308147863,
23
+ "qwen_2_5_7b_longfact_test/entity_tokens_auc": 0.8382365663877721,
24
+ "qwen_2_5_7b_longfact_test/entity_tokens_optimal_threshold": 0.13589015151515155,
25
+ "qwen_2_5_7b_longfact_test/entity_tokens_threshold_optimized_accuracy": 0.7659166021313031,
26
+ "qwen_2_5_7b_longfact_test/entity_tokens_recall_at_0.1_fpr": 0.5349468744232645,
27
+ "qwen_2_5_7b_longfact_test/entity_tokens_true_positive_count": 37929,
28
+ "qwen_2_5_7b_longfact_test/entity_tokens_true_negative_count": 53752,
29
+ "qwen_2_5_7b_longfact_test/entity_tokens_pred_positive_count": 14416,
30
+ "qwen_2_5_7b_longfact_test/entity_tokens_pred_negative_count": 77265,
31
+ "qwen_2_5_7b_longfact_test/entity_tokens_total_samples": 91681,
32
+ "qwen_2_5_7b_longfact_test/entity_span_max_accuracy": 0.8037324823088664,
33
+ "qwen_2_5_7b_longfact_test/entity_span_max_precision": 0.7878114976560573,
34
+ "qwen_2_5_7b_longfact_test/entity_span_max_recall": 0.6185586981790003,
35
+ "qwen_2_5_7b_longfact_test/entity_span_max_f1": 0.6930005425935974,
36
+ "qwen_2_5_7b_longfact_test/entity_span_max_auc": 0.8837167661414778,
37
+ "qwen_2_5_7b_longfact_test/entity_span_max_optimal_threshold": 0.437204071969697,
38
+ "qwen_2_5_7b_longfact_test/entity_span_max_threshold_optimized_accuracy": 0.8098376578326627,
39
+ "qwen_2_5_7b_longfact_test/entity_span_max_recall_at_0.1_fpr": 0.6375435877566834,
40
+ "qwen_2_5_7b_longfact_test/entity_span_max_true_positive_count": 5162,
41
+ "qwen_2_5_7b_longfact_test/entity_span_max_true_negative_count": 9252,
42
+ "qwen_2_5_7b_longfact_test/entity_span_max_pred_positive_count": 4053,
43
+ "qwen_2_5_7b_longfact_test/entity_span_max_pred_negative_count": 10361,
44
+ "qwen_2_5_7b_longfact_test/entity_span_max_total_samples": 14414,
45
+ "epoch": 1.0,
46
+ "qwen_2_5_7b_longfact_augmented_test/lm_loss": 0.31472883214076014,
47
+ "qwen_2_5_7b_longfact_augmented_test/probe_loss": 0.9969771905236889,
48
+ "qwen_2_5_7b_longfact_augmented_test/sparsity": 0.09702653702202062,
49
+ "qwen_2_5_7b_longfact_augmented_test/probe_threshold": 0.5,
50
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_accuracy": 0.9277186886236591,
51
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_precision": 0.3952882858072754,
52
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_recall": 0.3999442727595116,
53
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_f1": 0.39760264917339244,
54
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_auc": 0.8851051986364417,
55
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_optimal_threshold": 0.8589409722222228,
56
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_threshold_optimized_accuracy": 0.9409397189907841,
57
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_recall_at_0.1_fpr": 0.6396727291149501,
58
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_true_positive_count": 39478,
59
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_true_negative_count": 622422,
60
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_pred_positive_count": 39943,
61
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_pred_negative_count": 621957,
62
+ "qwen_2_5_7b_longfact_augmented_test/all_tokens_total_samples": 661900,
63
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_accuracy": 0.6896416008658648,
64
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_precision": 0.9108688127379716,
65
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_recall": 0.3999442727595116,
66
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_f1": 0.5558332746602831,
67
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_auc": 0.8556854617116961,
68
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_optimal_threshold": 0.12335759943181823,
69
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_threshold_optimized_accuracy": 0.7739527218163481,
70
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_recall_at_0.1_fpr": 0.5913420132732154,
71
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_true_positive_count": 39478,
72
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_true_negative_count": 41828,
73
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_pred_positive_count": 17334,
74
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_pred_negative_count": 63972,
75
+ "qwen_2_5_7b_longfact_augmented_test/entity_tokens_total_samples": 81306,
76
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_accuracy": 0.8000595356221473,
77
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_precision": 0.8051090700344432,
78
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_recall": 0.6773726153103116,
79
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_f1": 0.7357377049180328,
80
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_auc": 0.8888790523144552,
81
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_optimal_threshold": 0.4113794191919192,
82
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_threshold_optimized_accuracy": 0.8064100019845207,
83
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_recall_at_0.1_fpr": 0.644771794252596,
84
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_true_positive_count": 4141,
85
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_true_negative_count": 5937,
86
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_pred_positive_count": 3484,
87
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_pred_negative_count": 6594,
88
+ "qwen_2_5_7b_longfact_augmented_test/entity_span_max_total_samples": 10078,
89
+ "llama3_1_8b_trivia_qa_test/lm_loss": 1.2490080514870792,
90
+ "llama3_1_8b_trivia_qa_test/probe_loss": 0.888429161903274,
91
+ "llama3_1_8b_trivia_qa_test/sparsity": 0.19553201752473634,
92
+ "llama3_1_8b_trivia_qa_test/probe_threshold": 0.5,
93
+ "llama3_1_8b_trivia_qa_test/all_tokens_accuracy": 0.8640181090479627,
94
+ "llama3_1_8b_trivia_qa_test/all_tokens_precision": 0.19581519109820997,
95
+ "llama3_1_8b_trivia_qa_test/all_tokens_recall": 0.49662576687116566,
96
+ "llama3_1_8b_trivia_qa_test/all_tokens_f1": 0.28088133240804997,
97
+ "llama3_1_8b_trivia_qa_test/all_tokens_auc": 0.8182136579829025,
98
+ "llama3_1_8b_trivia_qa_test/all_tokens_optimal_threshold": 0.9733664772727275,
99
+ "llama3_1_8b_trivia_qa_test/all_tokens_threshold_optimized_accuracy": 0.9466734466242372,
100
+ "llama3_1_8b_trivia_qa_test/all_tokens_recall_at_0.1_fpr": 0.4634969325153374,
101
+ "llama3_1_8b_trivia_qa_test/all_tokens_true_positive_count": 3260,
102
+ "llama3_1_8b_trivia_qa_test/all_tokens_true_negative_count": 57704,
103
+ "llama3_1_8b_trivia_qa_test/all_tokens_pred_positive_count": 8268,
104
+ "llama3_1_8b_trivia_qa_test/all_tokens_pred_negative_count": 52696,
105
+ "llama3_1_8b_trivia_qa_test/all_tokens_total_samples": 60964,
106
+ "llama3_1_8b_trivia_qa_test/entity_tokens_accuracy": 0.6924650528469144,
107
+ "llama3_1_8b_trivia_qa_test/entity_tokens_precision": 0.9085297418630752,
108
+ "llama3_1_8b_trivia_qa_test/entity_tokens_recall": 0.49662576687116566,
109
+ "llama3_1_8b_trivia_qa_test/entity_tokens_f1": 0.6422054740182467,
110
+ "llama3_1_8b_trivia_qa_test/entity_tokens_auc": 0.8600766753457099,
111
+ "llama3_1_8b_trivia_qa_test/entity_tokens_optimal_threshold": 0.12021583017676771,
112
+ "llama3_1_8b_trivia_qa_test/entity_tokens_threshold_optimized_accuracy": 0.7877599727241732,
113
+ "llama3_1_8b_trivia_qa_test/entity_tokens_recall_at_0.1_fpr": 0.6058282208588958,
114
+ "llama3_1_8b_trivia_qa_test/entity_tokens_true_positive_count": 3260,
115
+ "llama3_1_8b_trivia_qa_test/entity_tokens_true_negative_count": 2606,
116
+ "llama3_1_8b_trivia_qa_test/entity_tokens_pred_positive_count": 1782,
117
+ "llama3_1_8b_trivia_qa_test/entity_tokens_pred_negative_count": 4084,
118
+ "llama3_1_8b_trivia_qa_test/entity_tokens_total_samples": 5866,
119
+ "llama3_1_8b_trivia_qa_test/entity_span_max_accuracy": 0.8570709612481128,
120
+ "llama3_1_8b_trivia_qa_test/entity_span_max_precision": 0.8701570680628272,
121
+ "llama3_1_8b_trivia_qa_test/entity_span_max_recall": 0.8385469223007064,
122
+ "llama3_1_8b_trivia_qa_test/entity_span_max_f1": 0.8540596094552929,
123
+ "llama3_1_8b_trivia_qa_test/entity_span_max_auc": 0.9361077002257263,
124
+ "llama3_1_8b_trivia_qa_test/entity_span_max_optimal_threshold": 0.5555555555555558,
125
+ "llama3_1_8b_trivia_qa_test/entity_span_max_threshold_optimized_accuracy": 0.8580775037745345,
126
+ "llama3_1_8b_trivia_qa_test/entity_span_max_recall_at_0.1_fpr": 0.8173562058526741,
127
+ "llama3_1_8b_trivia_qa_test/entity_span_max_true_positive_count": 991,
128
+ "llama3_1_8b_trivia_qa_test/entity_span_max_true_negative_count": 996,
129
+ "llama3_1_8b_trivia_qa_test/entity_span_max_pred_positive_count": 955,
130
+ "llama3_1_8b_trivia_qa_test/entity_span_max_pred_negative_count": 1032,
131
+ "llama3_1_8b_trivia_qa_test/entity_span_max_total_samples": 1987
132
+ }
qwen2_5_7b_lora_lambda_kl_0_05/llama3_1_8b_trivia_qa_test_roc_curves.png ADDED

Git LFS Details

  • SHA256: 5946b7028c42a5fe170b2b2b8f0a2fd7756228af4b4250f4da191c4627350acc
  • Pointer size: 131 Bytes
  • Size of remote file: 339 kB
qwen2_5_7b_lora_lambda_kl_0_05/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
qwen2_5_7b_lora_lambda_kl_0_05/probe_config.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "target_layer_name": "Qwen2DecoderLayer",
3
+ "layer_idx": 26,
4
+ "hidden_size": 3584
5
+ }
qwen2_5_7b_lora_lambda_kl_0_05/probe_head.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:caea04908e861b2daf2a248a450ac9d44860899b62bd776a17c38f6bbaf5175f
3
+ size 8682
qwen2_5_7b_lora_lambda_kl_0_05/qwen_2_5_7b_longfact_augmented_test_roc_curves.png ADDED

Git LFS Details

  • SHA256: e9d41fb1fa105129e22e9c839ea347a1816519de9f16323dbcd5d61e590875af
  • Pointer size: 131 Bytes
  • Size of remote file: 345 kB
qwen2_5_7b_lora_lambda_kl_0_05/qwen_2_5_7b_longfact_test_roc_curves.png ADDED

Git LFS Details

  • SHA256: 1c97aecf7f3ba3f8a6f22da2828ed06c790e320480087211f60f4ad8732ddd11
  • Pointer size: 131 Bytes
  • Size of remote file: 344 kB
qwen2_5_7b_lora_lambda_kl_0_05/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": "<|im_end|>",
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
+ }
qwen2_5_7b_lora_lambda_kl_0_05/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1e72410ddfe59d4ec8aeea923685710c76492112d74f5d127b86e2d08d65a3a4
3
+ size 11422174
qwen2_5_7b_lora_lambda_kl_0_05/tokenizer_config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ },
182
+ "additional_special_tokens": [
183
+ "<|im_start|>",
184
+ "<|im_end|>",
185
+ "<|object_ref_start|>",
186
+ "<|object_ref_end|>",
187
+ "<|box_start|>",
188
+ "<|box_end|>",
189
+ "<|quad_start|>",
190
+ "<|quad_end|>",
191
+ "<|vision_start|>",
192
+ "<|vision_end|>",
193
+ "<|vision_pad|>",
194
+ "<|image_pad|>",
195
+ "<|video_pad|>"
196
+ ],
197
+ "bos_token": null,
198
+ "chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0]['role'] == 'system' %}\n {{- messages[0]['content'] }}\n {%- else %}\n {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n {%- endif %}\n {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0]['role'] == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n {%- else %}\n {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {{- '<|im_start|>' + message.role }}\n {%- if message.content %}\n {{- '\\n' + message.content }}\n {%- endif %}\n {%- for tool_call in message.tool_calls %}\n {%- if tool_call.function is defined %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '\\n<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {{- tool_call.arguments | tojson }}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- message.content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
199
+ "clean_up_tokenization_spaces": false,
200
+ "eos_token": "<|im_end|>",
201
+ "errors": "replace",
202
+ "extra_special_tokens": {},
203
+ "model_max_length": 131072,
204
+ "pad_token": "<|endoftext|>",
205
+ "split_special_tokens": false,
206
+ "tokenizer_class": "Qwen2Tokenizer",
207
+ "unk_token": null
208
+ }
qwen2_5_7b_lora_lambda_kl_0_05/training_config.json ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "wandb_project": "probe-training",
3
+ "model_name": "Qwen/Qwen2.5-7B-Instruct",
4
+ "upload_to_hf": true,
5
+ "hf_repo_id": "andyrdt/hallucination-probes",
6
+ "save_evaluation_metrics": true,
7
+ "evaluation_output_dir": null,
8
+ "probe_id": "qwen2_5_7b_lora_lambda_kl=0.5",
9
+ "probe_dir": "/root/git/hallucination_detection/value_head_probes/qwen2_5_7b_lora_lambda_kl=0.5",
10
+ "layer": 26,
11
+ "lora_r": 16,
12
+ "lora_alpha": 32,
13
+ "lora_dropout": 0.05,
14
+ "lora_layers": [
15
+ 0,
16
+ 1,
17
+ 2,
18
+ 3,
19
+ 4,
20
+ 5,
21
+ 6,
22
+ 7,
23
+ 8,
24
+ 9,
25
+ 10,
26
+ 11,
27
+ 12,
28
+ 13,
29
+ 14,
30
+ 15,
31
+ 16,
32
+ 17,
33
+ 18,
34
+ 19,
35
+ 20,
36
+ 21,
37
+ 22,
38
+ 23,
39
+ 24,
40
+ 25,
41
+ 26
42
+ ],
43
+ "probe_threshold": 0.5,
44
+ "load_from_hf": false,
45
+ "load_from_disk": false,
46
+ "train_data_config_path": "experiments/training/qwen-2.5-7b/train_data.yaml",
47
+ "eval_data_config_path": "experiments/training/qwen-2.5-7b/eval_data.yaml",
48
+ "cache_data_config_path": null,
49
+ "high_loss_threshold": null,
50
+ "lambda_lm": 0.0,
51
+ "lambda_kl": 0.5,
52
+ "anneal_max_aggr": true,
53
+ "anneal_warmup": 1.0,
54
+ "learning_rate": 5e-05,
55
+ "probe_head_lr": 0.0005,
56
+ "lora_lr": 5e-05,
57
+ "sparsity_penalty_weight": null,
58
+ "use_focal_loss": false,
59
+ "focal_gamma": 2.0,
60
+ "evaluate_only": false,
61
+ "max_steps": -1,
62
+ "num_train_epochs": 1,
63
+ "per_device_train_batch_size": 4,
64
+ "per_device_eval_batch_size": 4,
65
+ "logging_steps": 10,
66
+ "eval_steps": null,
67
+ "cache_steps": null,
68
+ "evaluation_strategy": "no",
69
+ "seed": 42,
70
+ "save_roc_curves": true,
71
+ "dump_raw_eval_results": false,
72
+ "load_in_8bit": false,
73
+ "load_in_4bit": false,
74
+ "enable_gradient_checkpointing": true,
75
+ "gradient_accumulation_steps": 2,
76
+ "max_grad_norm": 1.0
77
+ }
qwen2_5_7b_lora_lambda_kl_0_05/vocab.json ADDED
The diff for this file is too large to render. See raw diff