vec2seq-llama2-reviewer-checkpoints / metadata /reviewer_evaluations.csv
vec2seq-artifact's picture
Add files using upload-large-folder tool
b49ab8d verified
Raw
History Blame Contribute Delete
21.2 kB
model_family,canonical_train_run_id,hub_slug,release_condition,evaluation_id,logical_evaluation_key,dataset,dataset_path,dataset_sha256,num_examples,t_tokens,obs_mode,latent_pooler_kind,prefix_component_mode,selected_checkpoint_step,selected_checkpoint_sha256,source_evaluation_protocol_namespace,reviewer_rerun_protocol_namespace,source_runtime_profile,source_slurm_partition,token_acc,exact_match,token_accuracy_definition,exact_match_definition,source_result_path,source_result_sha256,source_predictions_path,source_predictions_sha256,source_reconstructions_path,source_reconstructions_sha256,statistical_weight
llama,validation_exact_v1/experiments/capacity_study/train/fixed_linear/seed1234_train_87dcc3ad6872,llama-capacity-fixed-linear-t8-s1234,capacity_fixed_linear,llama_d9412291aa999b5a,"{""canonical_train_run_id"":""validation_exact_v1/experiments/capacity_study/train/fixed_linear/seed1234_train_87dcc3ad6872"",""dataset_id"":""data/test_sets/wikipedia_1000.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""capacity_study"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wikipedia,data/test_sets/wikipedia_1000.txt,2219cac7037e5f1c1abbf695215139207fd3900561098d9744eb0c49566c764a,1000,8,hidden_seq,fixed_linear,full,9000,5d7cf16dc5002b81ca5df87ba88e0edf9c45772c86d4ea492fee90fec22d9ece,cache_mask_modes_v1,cache_mask_modes_v1,default,ampere24,0.888875,0.623,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_d9412291aa999b5a/eval_results.json,53cbf89575fd3dd07f4ca720eed52dba6d3e5cefad9100933cb90e101b50e911,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_d9412291aa999b5a/eval_predictions.jsonl,2d6d776065f22cea8dbf81fb2a74354457a8da818c3ac54de35c8af0851d0854,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_d9412291aa999b5a/eval_reconstructions.csv,541e2b00aedb83a6330a8cd89598b16b1ba8144a402e7f777134baa90a252354,0
llama,validation_exact_v1/experiments/capacity_study/train/fixed_linear/seed1234_train_87dcc3ad6872,llama-capacity-fixed-linear-t8-s1234,capacity_fixed_linear,llama_5808140771f2cf7b,"{""canonical_train_run_id"":""validation_exact_v1/experiments/capacity_study/train/fixed_linear/seed1234_train_87dcc3ad6872"",""dataset_id"":""data/test_sets/wizardlm_217.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""capacity_study"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wizardlm,data/test_sets/wizardlm_217.txt,be90d3449149ff27288d513ee32d9826929c6be9415cf9207c918e7f160aa86b,217,8,hidden_seq,fixed_linear,full,9000,5d7cf16dc5002b81ca5df87ba88e0edf9c45772c86d4ea492fee90fec22d9ece,cache_mask_modes_v1,cache_mask_modes_v1,default,ampere24,0.8932411674347158,0.6820276497695853,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_5808140771f2cf7b/eval_results.json,9c34a63900b187bc8fff41cfff1ff9605a62a99d9a8dffa1afff63be69799494,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_5808140771f2cf7b/eval_predictions.jsonl,b36db04b649b6bad0ad0fef2ff3f2ddb4b0ec9184ebddb224d9c84f3c18f8302,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_5808140771f2cf7b/eval_reconstructions.csv,17b5952d7bfc323dddc9bb3f0fa018f1df05d4562e8d9b7a438739f116310196,0
llama,validation_exact_v1/experiments/capacity_study/train/fixed_residual_mlp/seed1234_train_b571d6924f6f,llama-capacity-fixed-residual-mlp-t8-s1234,capacity_fixed_residual_mlp,llama_41da95874044e917,"{""canonical_train_run_id"":""validation_exact_v1/experiments/capacity_study/train/fixed_residual_mlp/seed1234_train_b571d6924f6f"",""dataset_id"":""data/test_sets/wikipedia_1000.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""capacity_study"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wikipedia,data/test_sets/wikipedia_1000.txt,2219cac7037e5f1c1abbf695215139207fd3900561098d9744eb0c49566c764a,1000,8,hidden_seq,fixed_residual_mlp,full,10000,1c55e3249738c9abd322e6257d0457fda540a9e543a23666284436f92171d722,cache_mask_modes_v1,cache_mask_modes_v1,default,ampere80,0.9305,0.739,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_41da95874044e917/eval_results.json,20a9c5551701ee63c411c4283a5386096ee9046ac3d66a98a10c927cb7c3d0e0,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_41da95874044e917/eval_predictions.jsonl,94d97d8801d5d5d8e8918c8a5c34399efd420d8e313e2d5db17d996be4418ed2,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_41da95874044e917/eval_reconstructions.csv,be7410150d2a9b805cdf808c10a80d2c71b31b420909c9514593d3d98918b0b0,0
llama,validation_exact_v1/experiments/capacity_study/train/fixed_residual_mlp/seed1234_train_b571d6924f6f,llama-capacity-fixed-residual-mlp-t8-s1234,capacity_fixed_residual_mlp,llama_06091ba10d940a23,"{""canonical_train_run_id"":""validation_exact_v1/experiments/capacity_study/train/fixed_residual_mlp/seed1234_train_b571d6924f6f"",""dataset_id"":""data/test_sets/wizardlm_217.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""capacity_study"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wizardlm,data/test_sets/wizardlm_217.txt,be90d3449149ff27288d513ee32d9826929c6be9415cf9207c918e7f160aa86b,217,8,hidden_seq,fixed_residual_mlp,full,10000,1c55e3249738c9abd322e6257d0457fda540a9e543a23666284436f92171d722,cache_mask_modes_v1,cache_mask_modes_v1,default,ampere80,0.9312596006144394,0.7695852534562212,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_06091ba10d940a23/eval_results.json,e6219060279e4388ecb5615a05e05c99bb5bc0968bbb7614777db72e5166f80e,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_06091ba10d940a23/eval_predictions.jsonl,fc1aadb80bba75617d7fc9a9deda10b6e926163a69c687604f85046a8ae4c9f5,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_06091ba10d940a23/eval_reconstructions.csv,fbb3aeb0fce994227b4e868e9f91d7ddc60ff314d05c4d4ccd0c56248ec3e118,0
llama,by_representation/full_hidden_sequence/length_scaling/train/eosfix_v2/eosfix_v2_length_scaling_hidden_sequence_t32_final_primary_matched_seed1234,llama-length-full-t32-s1234,full_sequence_t32,llama_58effbebec5e9d0a,"{""canonical_train_run_id"":""by_representation/full_hidden_sequence/length_scaling/train/eosfix_v2/eosfix_v2_length_scaling_hidden_sequence_t32_final_primary_matched_seed1234"",""dataset_id"":""data/test_sets/wikipedia_1000.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""length_scaling"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wikipedia,data/test_sets/wikipedia_1000.txt,2219cac7037e5f1c1abbf695215139207fd3900561098d9744eb0c49566c764a,1000,32,hidden_seq,attention,full,18000,cc6e24e9e24950f4ebd856660de8e3125a5f1d49e3619d761bf22efac9ea2025,cache_mask_v1,cache_mask_v1,default,hopper96,0.885053587013192,0.329,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_58effbebec5e9d0a/eval_results.json,932373fbfd8d99e153d36f6d1f583b9c12e6c998f69add422314cb44fd85f31c,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_58effbebec5e9d0a/eval_predictions.jsonl,495f448ec85ae2d39166e1cf7ec9c50a56b51e0b95c38ae204d7b4f0a9ae2766,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_58effbebec5e9d0a/eval_reconstructions.csv,5a7d3421fdb6a4df3bb4362684ae96b549bd41218bf334b8ecb1ca9da6e9a996,0
llama,by_representation/full_hidden_sequence/length_scaling/train/eosfix_v2/eosfix_v2_length_scaling_hidden_sequence_t32_final_primary_matched_seed1234,llama-length-full-t32-s1234,full_sequence_t32,llama_43408cfe02b72249,"{""canonical_train_run_id"":""by_representation/full_hidden_sequence/length_scaling/train/eosfix_v2/eosfix_v2_length_scaling_hidden_sequence_t32_final_primary_matched_seed1234"",""dataset_id"":""data/test_sets/wizardlm_217.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""length_scaling"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wizardlm,data/test_sets/wizardlm_217.txt,be90d3449149ff27288d513ee32d9826929c6be9415cf9207c918e7f160aa86b,217,32,hidden_seq,attention,full,18000,cc6e24e9e24950f4ebd856660de8e3125a5f1d49e3619d761bf22efac9ea2025,cache_mask_v1,cache_mask_v1,default,ampere80,0.9302733571929775,0.5345622119815668,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_43408cfe02b72249/eval_results.json,4b88b4da49ea92c4f3da95251ca7625fdc944af689623ba6c0309191a88ae9f1,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_43408cfe02b72249/eval_predictions.jsonl,e1bf4ea2181058661b982c405a0d7b4a79d7796ec247468b722d2b5723c3aff9,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_43408cfe02b72249/eval_reconstructions.csv,424088fb0405504a04d8b62e7d627c84f828bd46f5caae4cb8363e0bfa7f28cf,0
llama,by_representation/full_hidden_sequence/prefix_components/train/eosfix_v2/eosfix_v2_prefix_components_hidden_sequence_t8_final_prefix_direct_only_seed1234,llama-prefix-direct-only-t8-s1234,prefix_direct_only,llama_3dcaf8f12f23f208,"{""canonical_train_run_id"":""by_representation/full_hidden_sequence/prefix_components/train/eosfix_v2/eosfix_v2_prefix_components_hidden_sequence_t8_final_prefix_direct_only_seed1234"",""dataset_id"":""data/test_sets/wikipedia_1000.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""prefix_components"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wikipedia,data/test_sets/wikipedia_1000.txt,2219cac7037e5f1c1abbf695215139207fd3900561098d9744eb0c49566c764a,1000,8,hidden_seq,attention,direct_only,11000,e748defca8c41f4919a75efec86ca5bb860ed896078e96055de8813bae718445,cache_mask_modes_v1,cache_mask_modes_v1,default,ampere40,0.850875,0.595,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_3dcaf8f12f23f208/eval_results.json,1f47d24e9ce2da24561909b30dacf6825a2779761c0fea1b33cb9191bfd85b0d,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_3dcaf8f12f23f208/eval_predictions.jsonl,5920b7827a8fb4514462677d968c2feaeec21ca941244d54367504ce0ff3aedc,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_3dcaf8f12f23f208/eval_reconstructions.csv,10b3292f44d75d3b3a6b204822ba576f7629abcd2c892e7f3b83adfbb22d465a,0
llama,by_representation/full_hidden_sequence/prefix_components/train/eosfix_v2/eosfix_v2_prefix_components_hidden_sequence_t8_final_prefix_direct_only_seed1234,llama-prefix-direct-only-t8-s1234,prefix_direct_only,llama_ce169c6d38c3137c,"{""canonical_train_run_id"":""by_representation/full_hidden_sequence/prefix_components/train/eosfix_v2/eosfix_v2_prefix_components_hidden_sequence_t8_final_prefix_direct_only_seed1234"",""dataset_id"":""data/test_sets/wizardlm_217.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""prefix_components"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wizardlm,data/test_sets/wizardlm_217.txt,be90d3449149ff27288d513ee32d9826929c6be9415cf9207c918e7f160aa86b,217,8,hidden_seq,attention,direct_only,11000,e748defca8c41f4919a75efec86ca5bb860ed896078e96055de8813bae718445,cache_mask_modes_v1,cache_mask_modes_v1,default,ampere40,0.8974654377880185,0.6451612903225806,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_ce169c6d38c3137c/eval_results.json,840233d353ff62c1ebad1f90ba07d96eca72710946043f65dc0ed55ef9daf632,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_ce169c6d38c3137c/eval_predictions.jsonl,b3c2182322b38d87b64c190ee7d87dbcd50fd7ed0365ff563ad2fe4328bd1b96,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_ce169c6d38c3137c/eval_reconstructions.csv,e47bd928383f88c8464aa3a9f286b63aed8714d5b74b09c4ef538be95ce892b2,0
llama,by_representation/last_token/main_t8/train/eosfix_v2/eosfix_v2_primary_last_token_t8_seed1234_hopper96_20260621_031008,llama-primary-final-t8-s1234,primary_final_t8,llama_51c052a61e27f32e,"{""canonical_train_run_id"":""by_representation/last_token/main_t8/train/eosfix_v2/eosfix_v2_primary_last_token_t8_seed1234_hopper96_20260621_031008"",""dataset_id"":""data/wikipedia_api_test_20260612/test_1000.txt"",""decode_method"":""current"",""eval_mode"":""last_token"",""experiment_family"":""main_t8"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wikipedia,data/test_sets/wikipedia_1000.txt,2219cac7037e5f1c1abbf695215139207fd3900561098d9744eb0c49566c764a,1000,8,last_token,attention,full,19000,54b0ad7bdbeca96c859408439e98922ff38674f53ee40bc9b8d5285792d979a0,validation_exact_v1,cache_mask_v1,historical,,0.143625,0.011,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_51c052a61e27f32e/eval_results.json,f670faa9733b81718434b71aa5f84c75aad113ce9b15bf7eb35b80f63923d0f6,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_51c052a61e27f32e/eval_predictions.jsonl,6c2ebcc4eccb4aedb3bad2c1d9ef3f5a325abda2301b405a83d36222da3e03ce,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_51c052a61e27f32e/eval_reconstructions.csv,8ee8f808058323ecdf07904c2667612159e2f3cfea8cc3ba0d98657b858f4851,0
llama,by_representation/last_token/main_t8/train/eosfix_v2/eosfix_v2_primary_last_token_t8_seed1234_hopper96_20260621_031008,llama-primary-final-t8-s1234,primary_final_t8,llama_febf7e68751fb522,"{""canonical_train_run_id"":""by_representation/last_token/main_t8/train/eosfix_v2/eosfix_v2_primary_last_token_t8_seed1234_hopper96_20260621_031008"",""dataset_id"":""data/test_sets/wizardlm_217.txt"",""decode_method"":""current"",""eval_mode"":""last_token"",""experiment_family"":""main_t8"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wizardlm,data/test_sets/wizardlm_217.txt,be90d3449149ff27288d513ee32d9826929c6be9415cf9207c918e7f160aa86b,217,8,last_token,attention,full,19000,54b0ad7bdbeca96c859408439e98922ff38674f53ee40bc9b8d5285792d979a0,validation_exact_v1,cache_mask_v1,historical,,0.08352534562211981,0.009216589861751152,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_febf7e68751fb522/eval_results.json,29465b5becdd4d8de2fead174cef9e1fd7d2431d07dda25166633acfa5bcceb4,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_febf7e68751fb522/eval_predictions.jsonl,5944131be61d8535b7719388a5f83c8b195478b39cb3809d0b6d75dbf17d1e66,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_febf7e68751fb522/eval_reconstructions.csv,b70e7ce96d78470ba3b644342410fd9e35817fdd844be0a41fd6f8dbad685dea,0
llama,by_representation/full_hidden_sequence/main_t8/train/eosfix_v2/eosfix_v2_main_t8_hidden_sequence_t8_final_primary_matched_seed1234,llama-primary-full-t8-s1234,primary_full_t8,llama_7d165e6c37324510,"{""canonical_train_run_id"":""by_representation/full_hidden_sequence/main_t8/train/eosfix_v2/eosfix_v2_main_t8_hidden_sequence_t8_final_primary_matched_seed1234"",""dataset_id"":""data/test_sets/wikipedia_1000.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""main_t8"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wikipedia,data/test_sets/wikipedia_1000.txt,2219cac7037e5f1c1abbf695215139207fd3900561098d9744eb0c49566c764a,1000,8,hidden_seq,attention,full,23000,acc1715b3d6797472ef7460c6ff30bf21f69d0ab33c85b66598eafa3897fdc39,validation_exact_v1,cache_mask_v1,historical,,0.922125,0.721,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_7d165e6c37324510/eval_results.json,befb46277b94b978f2c77a53c72ffe383be714a614375768faca4943a33f3396,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_7d165e6c37324510/eval_predictions.jsonl,c89fcb95ec17af68848aec86b375459c6cfd21a60ac1aa52b64446ecfdf199c0,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_7d165e6c37324510/eval_reconstructions.csv,3d53667bbb1d10f80700ef56c5d2ac258a9ecab823349c38c4156a8f0146cf92,0
llama,by_representation/full_hidden_sequence/main_t8/train/eosfix_v2/eosfix_v2_main_t8_hidden_sequence_t8_final_primary_matched_seed1234,llama-primary-full-t8-s1234,primary_full_t8,llama_88a5815b2f5cf117,"{""canonical_train_run_id"":""by_representation/full_hidden_sequence/main_t8/train/eosfix_v2/eosfix_v2_main_t8_hidden_sequence_t8_final_primary_matched_seed1234"",""dataset_id"":""data/test_sets/wizardlm_217.txt"",""decode_method"":""current"",""eval_mode"":""hidden_seq"",""experiment_family"":""main_t8"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wizardlm,data/test_sets/wizardlm_217.txt,be90d3449149ff27288d513ee32d9826929c6be9415cf9207c918e7f160aa86b,217,8,hidden_seq,attention,full,23000,acc1715b3d6797472ef7460c6ff30bf21f69d0ab33c85b66598eafa3897fdc39,cache_mask_v1,cache_mask_v1,hopper140,hopper140,0.940284178187404,0.7741935483870968,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_88a5815b2f5cf117/eval_results.json,90d98b6d277a633cc4c8033f1461e14f1472584733fff25bc746fa728c9d4607,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_88a5815b2f5cf117/eval_predictions.jsonl,af52015c1b2854cb1fab2f66e657a3bab378bdb7af857ec8f138c8999b09f090,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_88a5815b2f5cf117/eval_reconstructions.csv,52614aa54a6b67ccf0d59bcd8cb9217c7363f13401efdc6e830300f755bd85cd,0
llama,by_representation/pooled/main_t8/train/eosfix_v2/eosfix_v2_main_t8_pooled_vector_t8_final_primary_matched_seed1234,llama-primary-pooled-t8-s1234,primary_pooled_t8,llama_3214459a4265d91c,"{""canonical_train_run_id"":""by_representation/pooled/main_t8/train/eosfix_v2/eosfix_v2_main_t8_pooled_vector_t8_final_primary_matched_seed1234"",""dataset_id"":""data/test_sets/wikipedia_1000.txt"",""decode_method"":""current"",""eval_mode"":""single_vector"",""experiment_family"":""main_t8"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wikipedia,data/test_sets/wikipedia_1000.txt,2219cac7037e5f1c1abbf695215139207fd3900561098d9744eb0c49566c764a,1000,8,single_vector,attention,full,15000,ce59e0cdafaa47d88071a5d2acb11c746f79830c31fe4894830435a62c7684f2,validation_exact_v1,cache_mask_v1,historical,,0.508875,0.164,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_3214459a4265d91c/eval_results.json,c6e98ffcaa2d6ec7142daf8c000cc5c8e8c9c5472d62d479474ea577ebe3397b,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_3214459a4265d91c/eval_predictions.jsonl,ab1da587a953bf5e4d3118bfef68c4c9ea604f3a47dc5e2be12a94458aac2ffd,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_3214459a4265d91c/eval_reconstructions.csv,7f7ac9b7ecb7af988348277eb46c62733f00aa27bea0b62c79967d6ad88a251f,0
llama,by_representation/pooled/main_t8/train/eosfix_v2/eosfix_v2_main_t8_pooled_vector_t8_final_primary_matched_seed1234,llama-primary-pooled-t8-s1234,primary_pooled_t8,llama_d3f17474835e1992,"{""canonical_train_run_id"":""by_representation/pooled/main_t8/train/eosfix_v2/eosfix_v2_main_t8_pooled_vector_t8_final_primary_matched_seed1234"",""dataset_id"":""data/test_sets/wizardlm_217.txt"",""decode_method"":""current"",""eval_mode"":""single_vector"",""experiment_family"":""main_t8"",""obs_control"":""correct"",""obs_control_seed"":""20260613""}",wizardlm,data/test_sets/wizardlm_217.txt,be90d3449149ff27288d513ee32d9826929c6be9415cf9207c918e7f160aa86b,217,8,single_vector,attention,full,15000,ce59e0cdafaa47d88071a5d2acb11c746f79830c31fe4894830435a62c7684f2,validation_exact_v1,cache_mask_v1,historical,,0.4998079877112135,0.20276497695852536,gold_length_normalized,retained_sequence_length_sensitive,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_d3f17474835e1992/eval_results.json,e2f5a6dec90c12827ae4c7c59632230c74648c5ff46f07d41d4ca835cdf8a825,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_d3f17474835e1992/eval_predictions.jsonl,64b670f0200247673e6e2312d7af7d503551cb704c73c40dc87d8ddced12ee33,Llama-2-7b-hf/experiments/evidence/cache_mask_v1/canonical_evaluations/llama_d3f17474835e1992/eval_reconstructions.csv,ff4f64f3f4179562e25cf6a3e39eb07e3eab5c1934a804824b47bafe6c9d052d,0