Add reports from Benchmarking paper "Benchmark Leakage in Large Language Models"

#27
Files changed (1) hide show
  1. contamination_report.csv +10 -0
contamination_report.csv CHANGED
@@ -168,6 +168,8 @@ gsm8k;;EleutherAI/llemma_7b;;model;;;0.15;data-based;https://openreview.net/foru
168
  gsm8k;;EleutherAI/proof-pile-2;;corpus;;;0.15;data-based;https://openreview.net/forum?id=4WnqRR915j;23
169
  gsm8k;;GPT-4;;model;100.0;;1.0;data-based;https://arxiv.org/abs/2303.08774;11
170
  gsm8k;;GPT-4;;model;79.00;;;model-based;https://arxiv.org/abs/2311.06233;8
 
 
171
 
172
  head_qa;en;EleutherAI/pile;;corpus;;;5.11;data-based;https://arxiv.org/abs/2310.20707;2
173
  head_qa;en;allenai/c4;;corpus;;;5.22;data-based;https://arxiv.org/abs/2310.20707;2
@@ -179,9 +181,17 @@ health_fact;;allenai/c4;;corpus;;;7.53;data-based;https://arxiv.org/abs/2310.207
179
  health_fact;;oscar-corpus/OSCAR-2301;;corpus;;;3.4;data-based;https://arxiv.org/abs/2310.20707;2
180
  health_fact;;togethercomputer/RedPajama-Data-V2;;corpus;;;18.7;data-based;https://arxiv.org/abs/2310.20707;2
181
 
 
 
182
  hendrycks/competition_math;;EleutherAI/llemma_34b;;model;;;7.72;data-based;https://openreview.net/forum?id=4WnqRR915j;23
183
  hendrycks/competition_math;;EleutherAI/llemma_7b;;model;;;7.72;data-based;https://openreview.net/forum?id=4WnqRR915j;23
184
  hendrycks/competition_math;;EleutherAI/proof-pile-2;;corpus;;;7.72;data-based;https://openreview.net/forum?id=4WnqRR915j;23
 
 
 
 
 
 
185
 
186
  hlgd;;EleutherAI/pile;;corpus;;;0.0;data-based;https://arxiv.org/abs/2310.20707;2
187
  hlgd;;allenai/c4;;corpus;;;0.0;data-based;https://arxiv.org/abs/2310.20707;2
 
168
  gsm8k;;EleutherAI/proof-pile-2;;corpus;;;0.15;data-based;https://openreview.net/forum?id=4WnqRR915j;23
169
  gsm8k;;GPT-4;;model;100.0;;1.0;data-based;https://arxiv.org/abs/2303.08774;11
170
  gsm8k;;GPT-4;;model;79.00;;;model-based;https://arxiv.org/abs/2311.06233;8
171
+ gsm8k;;Qwen/Qwen-14B;;model;0.5;;;model-based;https://arxiv.org/abs/2404.18824;27
172
+ gsm8k;;Qwen/Qwen-1_8B;;model;12.8;;0.075;model-based;https://arxiv.org/abs/2404.18824;27
173
 
174
  head_qa;en;EleutherAI/pile;;corpus;;;5.11;data-based;https://arxiv.org/abs/2310.20707;2
175
  head_qa;en;allenai/c4;;corpus;;;5.22;data-based;https://arxiv.org/abs/2310.20707;2
 
181
  health_fact;;oscar-corpus/OSCAR-2301;;corpus;;;3.4;data-based;https://arxiv.org/abs/2310.20707;2
182
  health_fact;;togethercomputer/RedPajama-Data-V2;;corpus;;;18.7;data-based;https://arxiv.org/abs/2310.20707;2
183
 
184
+ hendrycks/competition_math;;BAAI/Aquila2-34B;;model;3.366;;1.166;model-based;https://arxiv.org/abs/2404.18824;27
185
+ hendrycks/competition_math;;BAAI/Aquila2-7B;;model;1;;0.133;model-based;https://arxiv.org/abs/2404.18824;27
186
  hendrycks/competition_math;;EleutherAI/llemma_34b;;model;;;7.72;data-based;https://openreview.net/forum?id=4WnqRR915j;23
187
  hendrycks/competition_math;;EleutherAI/llemma_7b;;model;;;7.72;data-based;https://openreview.net/forum?id=4WnqRR915j;23
188
  hendrycks/competition_math;;EleutherAI/proof-pile-2;;corpus;;;7.72;data-based;https://openreview.net/forum?id=4WnqRR915j;23
189
+ hendrycks/competition_math;;Qwen/Qwen-14B;;model;1.766;;1.6;model-based;https://arxiv.org/abs/2404.18824;27
190
+ hendrycks/competition_math;;Qwen/Qwen-1_8B;;model;4.533;;1.70;model-based;https://arxiv.org/abs/2404.18824;27
191
+ hendrycks/competition_math;;Qwen/Qwen-7B;;model;1.266;;0.766;model-based;https://arxiv.org/abs/2404.18824;27
192
+ hendrycks/competition_math;;THUDM/chatglm3-6b;;model;0.70;;0.4;model-based;https://arxiv.org/abs/2404.18824;27
193
+ hendrycks/competition_math;;internlm/internlm2-20b;;model;4.733;;0.666;model-based;https://arxiv.org/abs/2404.18824;27
194
+ hendrycks/competition_math;;internlm/internlm2-7b;;model;3.033;;0.433;model-based;https://arxiv.org/abs/2404.18824;27
195
 
196
  hlgd;;EleutherAI/pile;;corpus;;;0.0;data-based;https://arxiv.org/abs/2310.20707;2
197
  hlgd;;allenai/c4;;corpus;;;0.0;data-based;https://arxiv.org/abs/2310.20707;2