{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/census/softmax","entry":"softmax","source":"Syntology differential census (groundwork/55, run v2_2026-09-22), per sample; not an archive number","census_date":"2026-09-22","battery_sha256":["b986f7e04d794a0d88ad4c5f32cf63ec3590b5deff0192150737bc5f1c0b4677"],"runner_sha256":["5a452d0e7c0da5b80771d1be2afe3572e253568cf5c5d59a0d08ebd663d00808"],"bucket_key":"positional (rank, kind, dtype) of each array argument; the argument name is not part of the key because the harness draws the shared array from (rank, kind) and casts it to the dtype, whatever the name","bucket_fields":["rank","kind","dtype"],"claim":"Implementations sharing this entry name were each run on one shared input fixed by the positional (rank, kind, dtype) of their array arguments (the bucket). A cluster is the set whose recorded output digest (sha256 of the output rounded to 6 decimals) is identical. Identical values to six decimals on the shared input are agreement on those inputs, not a statement about the whole domain and not a substitution claim.","n_implementations_compared":53,"n_papers":63,"n_buckets":8,"n_distinct_outputs":16,"n_class_bearing":0,"not_compared":{"not_run_on_shared_input":{"n":1,"by_error":{"ValueError":1}},"no_array_argument_ran_on_own_fixture_arguments_only":{"n":1,"n_papers":1,"recorded_shared_digest_equals_own_fixture_digest":1},"output_not_digested_non_numeric":{"n":0,"by_type":{}}},"withdrawn_excluded":0,"code_page":"/code/softmax","buckets":[{"bucket":[[2,"float","float32"]],"n_implementations_compared":25,"n_papers":25,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"95d860eeb84c6f36","size":10,"n_papers":10,"shape":[4,8],"dtype":"float32","type":"ndarray","finite":true,"max_difference_among_recorded_values":1.4901161193847656e-08,"members_with_recorded_values":10,"torch_reference_conventions_with_this_digest":[],"values":[0.5300593972206116,0.018851345404982567,0.1406695693731308,0.0437929704785347,0.03542829677462578,0.031183501705527306,0.045042190700769424,0.15497279167175293,0.09107176214456558,0.14144591987133026,0.0364365428686142,0.01129571720957756,0.04061903432011604,0.5527360439300537,0.03239447623491287,0.09400051832199097,0.13331103324890137,0.0246164258569479,0.19217392802238464,0.08310489356517792,0.12196025997400284,0.038302190601825714,0.07292425632476807,0.3336069881916046,0.037543922662734985,0.12830352783203125,0.21402700245380402,0.23012961447238922,0.024418439716100693,0.2974786162376404,0.022173603996634483,0.04592525586485863],"values_recorded":32,"members":[{"code_sha256_prefix":"08bc835660d4723f","path":"caliblab/calibrators/conformal_mass_threshold_calibrator.py","papers":["2505.15437"],"paper_pages":[{"arxiv_id":"2505.15437","page":"/paper/adaptive-temperature-scaling-with-conformal"}],"arg_sig_recorded":[["logits",2,"float","float32"]],"scalar_args":{"axis":"-1"},"class_bearing":false},{"code_sha256_prefix":"74138c11e6beb035","path":"mr_uplift/erupt.py","papers":["1705.08492"],"paper_pages":[{"arxiv_id":"1705.08492","page":"/paper/uplift-modeling-with-multiple-treatments-and"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"785d346b27b67e09","path":"utils.py","papers":["2205.12134"],"paper_pages":[{"arxiv_id":"2205.12134","page":"/paper/adversarial-attack-on-attackers-post-process"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{"axis":"1"},"class_bearing":false},{"code_sha256_prefix":"7c0d7e9e7a7f89cc","path":"src/benchmark/softmax.py","papers":["2403.14111"],"paper_pages":[{"arxiv_id":"2403.14111","page":"/paper/hetal-efficient-privacy-preserving-transfer"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"7ceb8f93a02b4ee3","path":"code/stac.py","papers":["2607.10804"],"paper_pages":[{"arxiv_id":"2607.10804","page":"/paper/arxiv-2607-10804"}],"arg_sig_recorded":[["Z",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"7ddc824e88e78020","path":"src/score_utils.py","papers":["1610.02136"],"paper_pages":[{"arxiv_id":"1610.02136","page":"/paper/a-baseline-for-detecting-misclassified-and"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"8d6c2bedc92a8075","path":"PACE/src/model.py","papers":["2406.12649"],"paper_pages":[{"arxiv_id":"2406.12649","page":"/paper/probabilistic-conceptual-explainers"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"a16c332f67a73877","path":"utils.py","papers":["2403.12729"],"paper_pages":[{"arxiv_id":"2403.12729","page":"/paper/posterior-uncertainty-quantification-in"}],"arg_sig_recorded":[["logp",2,"float","float32"]],"scalar_args":{"axis":"-1"},"class_bearing":false},{"code_sha256_prefix":"db83f3a3d4f7c604","path":"utils.py","papers":["1503.00269"],"paper_pages":[{"arxiv_id":"1503.00269","page":"/paper/contrastive-pessimistic-likelihood-estimation"}],"arg_sig_recorded":[["X",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"f89fc809b3a7824a","path":"src/grounding_evaluator.py","papers":["2209.14941"],"paper_pages":[{"arxiv_id":"2209.14941","page":"/paper/eda-explicit-text-decoupling-and-dense"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false}]},{"output_sha":"09f6f45b4d587d34","size":8,"n_papers":9,"shape":[4,8],"dtype":"float32","type":"ndarray","finite":true,"max_difference_among_recorded_values":5.960464477539063e-08,"members_with_recorded_values":8,"torch_reference_conventions_with_this_digest":["dim0"],"values":[0.6546289324760437,0.04725700989365578,0.2517068684101105,0.12098696082830429,0.16711440682411194,0.024717066437005997,0.26241371035575867,0.26335275173187256,0.18132950365543365,0.5716476440429688,0.1051105260848999,0.0503108911216259,0.3088924288749695,0.7063239812850952,0.3042650520801544,0.2575293779373169,0.11533434689044952,0.04322856664657593,0.2408859133720398,0.16083578765392303,0.4029989540576935,0.021267561241984367,0.2976192235946655,0.39713624119758606,0.04870721697807312,0.33786672353744507,0.40229666233062744,0.6678662896156311,0.12099418044090271,0.2476913034915924,0.1357019990682602,0.08198170363903046],"values_recorded":32,"members":[{"code_sha256_prefix":"cb5cbea0e8219ff5","path":"bert/evaluate.py","papers":["1301.3781","2004.08056"],"paper_pages":[{"arxiv_id":"1301.3781","page":"/paper/efficient-estimation-of-word-representations"},{"arxiv_id":"2004.08056","page":"/paper/dialogue-based-relation-extraction"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"083546dc5ac62494","path":"dreamerv2/goal_picker.py","papers":["2007.02832"],"paper_pages":[{"arxiv_id":"2007.02832","page":"/paper/maximum-entropy-gain-exploration-for-long"}],"arg_sig_recorded":[["X",2,"float","float32"]],"scalar_args":{"theta":"1.0","axis":"None"},"class_bearing":false},{"code_sha256_prefix":"51f9969762b91ea3","path":"train_model.py","papers":["1810.03947"],"paper_pages":[{"arxiv_id":"1810.03947","page":"/paper/texttovec-deep-contextualized-neural"}],"arg_sig_recorded":[["X",2,"float","float32"]],"scalar_args":{"theta":"1.0","axis":"None"},"class_bearing":false},{"code_sha256_prefix":"54fe686797cba4a9","path":"transformer.py","papers":["1706.03762"],"paper_pages":[{"arxiv_id":"1706.03762","page":"/paper/attention-is-all-you-need"}],"arg_sig_recorded":[["ys",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"81767458c6ff4e4d","path":"mDLAM.py","papers":["1811.04187"],"paper_pages":[{"arxiv_id":"1811.04187","page":"/paper/the-global-convergence-of-the-alternating"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"a49ec0b422e77ed3","path":"1_easyrun/RankC.py","papers":["2310.10378"],"paper_pages":[{"arxiv_id":"2310.10378","page":"/paper/cross-lingual-consistency-of-factual"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"b2d8244fa9563626","path":"al_neural_dialogue_train.py","papers":["1701.06547"],"paper_pages":[{"arxiv_id":"1701.06547","page":"/paper/adversarial-learning-for-neural-dialogue"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"cd98260d20abd01f","path":"test_demo.py","papers":["2006.04176"],"paper_pages":[{"arxiv_id":"2006.04176","page":"/paper/deep-active-inference-agents-using-monte"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{"temp":"1.0"},"class_bearing":false}]},{"output_sha":"ab42e90e344c9899","size":6,"n_papers":5,"shape":[4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":1.1920928955078125e-07,"members_with_recorded_values":6,"torch_reference_conventions_with_this_digest":["last_dim"],"values":[0.5300593376159668,0.018851343542337418,0.1406695693731308,0.0437929667532444,0.03542829304933548,0.031183499842882156,0.045042186975479126,0.15497279167175293,0.09107175469398499,0.14144591987133026,0.0364365391433239,0.011295716278254986,0.040619030594825745,0.5527359843254089,0.03239447623491287,0.09400051087141037,0.13331103324890137,0.0246164258569479,0.19217392802238464,0.08310489356517792,0.12196025997400284,0.038302190601825714,0.07292425632476807,0.3336069881916046,0.037543922662734985,0.12830352783203125,0.21402700245380402,0.23012961447238922,0.024418439716100693,0.2974786162376404,0.022173603996634483,0.04592525586485863],"values_recorded":32,"members":[{"code_sha256_prefix":"348d3e3d68c0c2dc","path":"soft_moe/soft_moe.py","papers":["2308.00951"],"paper_pages":[{"arxiv_id":"2308.00951","page":"/paper/from-sparse-to-soft-mixtures-of-experts"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{"dim":"1"},"class_bearing":false},{"code_sha256_prefix":"62d3b888dc672b51","path":"src/train_eval.py","papers":["2312.12736"],"paper_pages":[{"arxiv_id":"2312.12736","page":"/paper/learning-and-forgetting-unsafe-examples-in"}],"arg_sig_recorded":[["tensor",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"84195cc106a2908b","path":"soft_moe/soft_moe.py","papers":["2308.00951"],"paper_pages":[{"arxiv_id":"2308.00951","page":"/paper/from-sparse-to-soft-mixtures-of-experts"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{"dim":"1"},"class_bearing":false},{"code_sha256_prefix":"8825c7c596e27b6f","path":"lib/utils.py","papers":["1506.02640"],"paper_pages":[{"arxiv_id":"1506.02640","page":"/paper/you-only-look-once-unified-real-time-object"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"c42e5fa46bf385ab","path":"models/dgrmil.py","papers":["2407.03575"],"paper_pages":[{"arxiv_id":"2407.03575","page":"/paper/dgr-mil-exploring-diverse-global"}],"arg_sig_recorded":[["x",2,"float","float32"]],"scalar_args":{"dim":"1","onnx_trace":"False"},"class_bearing":false},{"code_sha256_prefix":"e3599c195e20c0fc","path":"src/models/programs.py","papers":["2306.01128"],"paper_pages":[{"arxiv_id":"2306.01128","page":"/paper/learning-transformer-programs-1"}],"arg_sig_recorded":[["X",2,"float","float32"]],"scalar_args":{"dim":"-1","tau":"1.0"},"class_bearing":false}]},{"output_sha":"410404bd78e6309e","size":1,"n_papers":1,"shape":[],"dtype":"float64","type":"float","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[3.4295716285705566],"values_recorded":1,"members":[{"code_sha256_prefix":"6a445568164614eb","path":"scripts/04_evaluate.py","papers":["2505.18513"],"paper_pages":[{"arxiv_id":"2505.18513","page":"/paper/enhancing-training-data-attribution-with"}],"arg_sig_recorded":[["a",2,"float","float32"]],"scalar_args":{},"class_bearing":false}]}]},{"bucket":[[1,"float","float64"]],"n_implementations_compared":14,"n_papers":23,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"aa13e79c00659056","size":11,"n_papers":19,"shape":[8],"dtype":"float64","type":"ndarray","finite":true,"max_difference_among_recorded_values":1.0930145677434666e-13,"members_with_recorded_values":11,"torch_reference_conventions_with_this_digest":[],"values":[0.05456670870405572,0.15766056874503237,0.33060521266653586,0.06287303808053657,0.1129626691483986,0.1420735223702735,0.026439859056393226,0.11281842122877413],"values_recorded":8,"members":[{"code_sha256_prefix":"19f01570b0a18e2b","path":"smoe/entrypoint/eval/eval_mmlu_moe_0.py","papers":["2009.03300","2505.24449","2404.02078","2601.16503"],"paper_pages":[{"arxiv_id":"2009.03300","page":"/paper/measuring-massive-multitask-language"},{"arxiv_id":"2505.24449","page":"/paper/when-large-multimodal-models-confront"},{"arxiv_id":"2404.02078","page":"/paper/advancing-llm-reasoning-generalists-with"},{"arxiv_id":"2601.16503","page":"/paper/arxiv-2601-16503"}],"arg_sig_recorded":[["x",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"7f9b15275ca5759c","path":"random_mutation_search.py","papers":["2103.15692","2401.12794","2510.05566","2402.14418"],"paper_pages":[{"arxiv_id":"2103.15692","page":"/paper/self-constructing-neural-networks-through"},{"arxiv_id":"2401.12794","page":"/paper/benchmarking-llms-via-uncertainty"},{"arxiv_id":"2510.05566","page":"/paper/arxiv-2510-05566"},{"arxiv_id":"2402.14418","page":"/paper/uncertainty-aware-evaluation-for-vision"}],"arg_sig_recorded":[["x",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"c0cf1196424d4f71","path":"word2vec.py","papers":["1411.2738","1906.05394","2506.14605"],"paper_pages":[{"arxiv_id":"1411.2738","page":"/paper/word2vec-parameter-learning-explained"},{"arxiv_id":"1906.05394","page":"/paper/neural-arabic-question-answering"},{"arxiv_id":"2506.14605","page":"/paper/unsupervised-imaging-inverse-problems-with"}],"arg_sig_recorded":[["x",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"9c4be464cbe63fba","path":"query_openai.py","papers":["2404.19509","2212.06801"],"paper_pages":[{"arxiv_id":"2404.19509","page":"/paper/do-large-language-models-understand"},{"arxiv_id":"2212.06801","page":"/paper/a-fine-grained-comparison-of-pragmatic"}],"arg_sig_recorded":[["x",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"08883768489fd226","path":"src/vlmeval/calculate_scores/map_predictions.py","papers":["2504.05457"],"paper_pages":[{"arxiv_id":"2504.05457","page":null}],"arg_sig_recorded":[["x",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"380f691b7059126a","path":"search.py","papers":["2502.04510"],"paper_pages":[{"arxiv_id":"2502.04510","page":"/paper/heterogeneous-swarms-jointly-optimizing-model"}],"arg_sig_recorded":[["probs",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"4471357a1f78d539","path":"src/compute_results.py","papers":["2104.00795"],"paper_pages":[{"arxiv_id":"2104.00795","page":"/paper/no-cost-likelihood-manipulation-at-test-time"}],"arg_sig_recorded":[["x",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"50d5258bd0761c1b","path":"mmd_dilated/mmd_dilated.py","papers":["2206.05825"],"paper_pages":[{"arxiv_id":"2206.05825","page":"/paper/a-unified-approach-to-reinforcement-learning"}],"arg_sig_recorded":[["x",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"54948e052031386e","path":"OpenAttack/attackers/genetic.py","papers":["2111.07970"],"paper_pages":[{"arxiv_id":"2111.07970","page":"/paper/triggerless-backdoor-attack-for-nlp-tasks"}],"arg_sig_recorded":[["inputs",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"885d6e6994f5e4bd","path":"src/cp_methods.py","papers":["2510.05566"],"paper_pages":[{"arxiv_id":"2510.05566","page":"/paper/arxiv-2510-05566"}],"arg_sig_recorded":[["x",1,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"ad039f77ef1e89fc","path":"toy/ARSM_Univariate.py","papers":["1905.01413"],"paper_pages":[{"arxiv_id":"1905.01413","page":"/paper/arsm-augment-reinforce-swap-merge-estimator"}],"arg_sig_recorded":[["phi",1,"float","float64"]],"scalar_args":{},"class_bearing":false}]},{"output_sha":"231a35f00878e97d","size":1,"n_papers":2,"shape":[8],"dtype":"float64","type":"ndarray","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.06465900216361926,0.15653991951848947,0.29014423522380844,0.07276282812226054,0.1185682254657247,0.14353246894510646,0.03535128015039555,0.11844204041059561],"values_recorded":8,"members":[{"code_sha256_prefix":"69e582d84850bb60","path":"MAB/train_mab_mo_multiple.py","papers":["2412.01572","2412.07618"],"paper_pages":[{"arxiv_id":"2412.01572","page":"/paper/mba-rag-a-bandit-approach-for-adaptive"},{"arxiv_id":"2412.07618","page":"/paper/adapting-to-non-stationary-environments-multi"}],"arg_sig_recorded":[["data",1,"float","float64"]],"scalar_args":{"tau":"1.2"},"class_bearing":false}]},{"output_sha":"0c77544de6fbb01e","size":1,"n_papers":1,"shape":[8],"dtype":"float64","type":"ndarray","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.08755765128661856,0.14883044079665628,0.21551875619834862,0.09398584421045786,0.12597886647674264,0.14128198806892042,0.06094804666985452,0.12589840629240112],"values_recorded":8,"members":[{"code_sha256_prefix":"3be33e47b8058815","path":"rl_algorithm/common/option_model.py","papers":["2310.03342"],"paper_pages":[{"arxiv_id":"2310.03342","page":"/paper/lesson-learning-to-integrate-exploration"}],"arg_sig_recorded":[["a",1,"float","float64"]],"scalar_args":{"ww":"0.5"},"class_bearing":false}]},{"output_sha":"92bf9c23a889056c","size":1,"n_papers":1,"shape":[8],"dtype":"float64","type":"ndarray","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.0989502939674489,0.14292775774831085,0.18474385676828106,0.10393075431538082,0.12733178694000524,0.13786309019909712,0.07697704833349563,0.12727541172798035],"values_recorded":8,"members":[{"code_sha256_prefix":"4a6f0cbbf7e6a8d9","path":"discrete_usher/clean_q_implementation.py","papers":["2207.01115"],"paper_pages":[{"arxiv_id":"2207.01115","page":"/paper/usher-unbiased-sampling-for-hindsight"}],"arg_sig_recorded":[["arr",1,"float","float64"]],"scalar_args":{"temp":"0.5"},"class_bearing":false}]}]},{"bucket":[[2,"float","float64"]],"n_implementations_compared":9,"n_papers":12,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"ab42e90e344c9899","size":6,"n_papers":7,"shape":[4,8],"dtype":"float64","type":"ndarray","finite":true,"max_difference_among_recorded_values":3.055178332544983e-11,"members_with_recorded_values":6,"torch_reference_conventions_with_this_digest":[],"values":[0.5300593666283869,0.018851341458504454,0.14066955557547944,0.04379296863982826,0.03542829649123797,0.03118349997228547,0.0450421811308727,0.15497279010340487,0.09107175306298512,0.14144590957871817,0.03643654172866276,0.011295716226771928,0.04061902883966086,0.5527360592405934,0.03239447103183824,0.09400052029076944,0.1333110249481478,0.024616426493662406,0.19217394235950472,0.08310489545827153,0.12196026128699544,0.03830219090945938,0.07292425058618603,0.33360700795777276,0.03754392718817198,0.1283035362723642,0.21402700561964594,0.2301296112978347,0.024418441682223627,0.2974786130350462,0.022173603092203197,0.045925261812510144],"values_recorded":32,"members":[{"code_sha256_prefix":"a8ae5c6e4eb58226","path":"metrics.py","papers":["2207.03036","2308.15074"],"paper_pages":[{"arxiv_id":"2207.03036","page":"/paper/not-all-models-are-equal-predicting-model"},{"arxiv_id":"2308.15074","page":"/paper/exploring-model-transferability-through-the"}],"arg_sig_recorded":[["X",2,"float","float64"]],"scalar_args":{"copy":"True"},"class_bearing":false},{"code_sha256_prefix":"475ce08eea0c1d56","path":"yolo_detection.py","papers":["1506.02640"],"paper_pages":[{"arxiv_id":"1506.02640","page":"/paper/you-only-look-once-unified-real-time-object"}],"arg_sig_recorded":[["val",2,"float","float64"]],"scalar_args":{"axis":"-1"},"class_bearing":false},{"code_sha256_prefix":"64e12b411d5ec4ca","path":"code/debias_utils.py","papers":["2309.03882"],"paper_pages":[{"arxiv_id":"2309.03882","page":"/paper/on-large-language-models-selection-bias-in"}],"arg_sig_recorded":[["x",2,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"652f07c5ff410713","path":"evaluation.py","papers":["1803.08024"],"paper_pages":[{"arxiv_id":"1803.08024","page":"/paper/stacked-cross-attention-for-image-text"}],"arg_sig_recorded":[["X",2,"float","float64"]],"scalar_args":{"axis":"1"},"class_bearing":false},{"code_sha256_prefix":"9374cbb282b14e3e","path":"a2/word2vec.py","papers":["1301.3781"],"paper_pages":[{"arxiv_id":"1301.3781","page":"/paper/efficient-estimation-of-word-representations"}],"arg_sig_recorded":[["x",2,"float","float64"]],"scalar_args":{},"class_bearing":false},{"code_sha256_prefix":"aa0acf499f826391","path":"logistic.py","papers":["1906.01827"],"paper_pages":[{"arxiv_id":"1906.01827","page":"/paper/data-sketching-for-faster-training-of-machine"}],"arg_sig_recorded":[["x",2,"float","float64"]],"scalar_args":{},"class_bearing":false}]},{"output_sha":"8f915a52849e3763","size":2,"n_papers":3,"shape":[4,8],"dtype":"float64","type":"ndarray","finite":true,"max_difference_among_recorded_values":0.0,"members_with_recorded_values":2,"torch_reference_conventions_with_this_digest":[],"values":[0.121484853105239,0.0043205583979967505,0.03224020057256578,0.010036955664216642,0.008119847824563283,0.007146978530984503,0.010323263963864446,0.03551837365081461,0.03365080109729546,0.05226393485549392,0.01346321748672779,0.004173740893465742,0.01500863675374092,0.20423468927785304,0.011969681758410606,0.034732973781226195,0.021403540503822066,0.003952251373968465,0.030854182995522547,0.013342774888264979,0.019581136618882444,0.00614954760744362,0.011708237572544355,0.05356174487414205,0.00903899616468144,0.030890086869930177,0.05152868726912817,0.05540551640948462,0.005878932153444322,0.07162031901515255,0.005338469582658274,0.011056868486471258],"values_recorded":32,"members":[{"code_sha256_prefix":"98bba89c2a19a3bb","path":"COMMON/src/lap_solvers/ILP.py","papers":["2212.04085","2312.02224"],"paper_pages":[{"arxiv_id":"2212.04085","page":"/paper/graph-matching-with-bi-level-noisy"},{"arxiv_id":"2312.02224","page":"/paper/tracing-hyperparameter-dependencies-for-model"}],"arg_sig_recorded":[["x",2,"float","float64"]],"scalar_args":{"axis":"None"},"class_bearing":false},{"code_sha256_prefix":"d8e37876ec1be74c","path":"generate_data.py","papers":["2111.02080"],"paper_pages":[{"arxiv_id":"2111.02080","page":"/paper/an-explanation-of-in-context-learning-as-1"}],"arg_sig_recorded":[["x",2,"float","float64"]],"scalar_args":{"temp":"1.0","axis":"None"},"class_bearing":false}]},{"output_sha":"09f6f45b4d587d34","size":1,"n_papers":2,"shape":[4,8],"dtype":"float64","type":"ndarray","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.6546289331468981,0.04725700680243002,0.25170688443267847,0.1209869590225412,0.16711441820397066,0.024717069356169234,0.2624137024293524,0.2633527394986028,0.18132950288689945,0.5716476662213358,0.10511052871387634,0.05031089459214579,0.3088924390366938,0.7063240721839678,0.3042650578460839,0.25752935329034343,0.11533435261633636,0.04322857206400423,0.24088591682420313,0.16083579649641688,0.40299896310101124,0.021267559999415663,0.2976192395242875,0.3971362085326518,0.04870721134986611,0.33786675491222995,0.40229667002924197,0.6678663498888961,0.12099417965832439,0.2476912984604473,0.13570200020027626,0.08198169867840198],"values_recorded":32,"members":[{"code_sha256_prefix":"91fe7a23c1c052b7","path":"image_classification/assistant.py","papers":["1804.08049","1905.03381"],"paper_pages":[{"arxiv_id":"1804.08049","page":"/paper/semi-supervised-user-geolocation-via-graph"},{"arxiv_id":"1905.03381","page":"/paper/190503381"}],"arg_sig_recorded":[["x",2,"float","float64"]],"scalar_args":{},"class_bearing":false}]}]},{"bucket":[[1,"float","float32"]],"n_implementations_compared":1,"n_papers":1,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"0c77544de6fbb01e","size":1,"n_papers":1,"shape":[8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.08755765110254288,0.14883044362068176,0.2155187577009201,0.09398584812879562,0.12597885727882385,0.14128199219703674,0.060948047786951065,0.12589840590953827],"values_recorded":8,"members":[{"code_sha256_prefix":"9559ea8ea49d3010","path":"src/models/_A2C_discrete.py","papers":["1602.01783"],"paper_pages":[{"arxiv_id":"1602.01783","page":"/paper/asynchronous-methods-for-deep-reinforcement"}],"arg_sig_recorded":[["z",1,"float","float32"]],"scalar_args":{"beta":"2.0"},"class_bearing":false}]}]},{"bucket":[[2,"float","float32"],[1,"float","float32"]],"n_implementations_compared":1,"n_papers":1,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"4992386467909321","size":1,"n_papers":1,"shape":[4,8],"dtype":"float32","type":"ndarray","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.26716217398643494,0.02745288796722889,0.4295685589313507,0.02543264627456665,0.036966465413570404,0.040922366082668304,0.011000217869877815,0.1614946573972702,0.036918941885232925,0.16567283868789673,0.0894920825958252,0.005276136100292206,0.03408806025981903,0.5834030508995056,0.0063630868680775166,0.07878582924604416,0.05244738236069679,0.0279819555580616,0.4580724537372589,0.03767220303416252,0.09933064877986908,0.03923437371850014,0.013901482336223125,0.2713595926761627,0.012942037545144558,0.12779000401496887,0.44700607657432556,0.0914054587483406,0.01742561347782612,0.2669955790042877,0.0037036556750535965,0.032731570303440094],"values_recorded":32,"members":[{"code_sha256_prefix":"1a0b58e25479df38","path":"abstention/calibration.py","papers":["1901.06852"],"paper_pages":[{"arxiv_id":"1901.06852","page":"/paper/calibration-with-bias-corrected-temperature"}],"arg_sig_recorded":[["preact",2,"float","float32"],["biases",1,"float","float32"]],"scalar_args":{"temp":"1.0"},"class_bearing":false}]}]},{"bucket":[[2,"float","float32"],[2,"float","float32"]],"n_implementations_compared":1,"n_papers":1,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"feda954aa726c243","size":1,"n_papers":1,"shape":[4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.997279703617096,-0.030177997425198555,0.06989430636167526,-0.031578946858644485,-0.033385034650564194,-0.033538755029439926,-0.031157491728663445,0.09266425669193268,0.03241967782378197,0.0930757150053978,-0.009929042309522629,-0.012153957039117813,-0.008040621876716614,0.8805655837059021,-0.011440825648605824,0.035503532737493515,0.03414646536111832,-0.1541173756122589,0.320355087518692,-0.1302245408296585,-0.010630586184561253,-0.17447568476200104,-0.15103623270988464,1.2659828662872314,-0.039581842720508575,0.07366415858268738,0.2680061161518097,0.3102910816669464,-0.03966313973069191,0.5022894144058228,-0.038850367069244385,-0.0361553356051445],"values_recorded":32,"members":[{"code_sha256_prefix":"2f6500993366c0f7","path":"graph_transformer/graph_transformer_model.py","papers":["2009.03509"],"paper_pages":[{"arxiv_id":"2009.03509","page":"/paper/masked-label-prediction-unified-massage"}],"arg_sig_recorded":[["x",2,"float","float32"],["adjacency",2,"float","float32"]],"scalar_args":{"dim":"-1"},"class_bearing":false}]}]},{"bucket":[[3,"float","float32"]],"n_implementations_compared":1,"n_papers":1,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"e53c681093b72098","size":1,"n_papers":1,"shape":[2,4,8],"dtype":"float32","type":"Tensor","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.04230783134698868,0.5732800364494324,0.13534824550151825,0.18657585978507996,0.264771431684494,0.031258177012205124,0.24162185192108154,0.061738405376672745,0.01216892246156931,0.10781250149011612,0.20391567051410675,0.11159101873636246,0.5255683064460754,0.11078701168298721,0.39512789249420166,0.13576802611351013,0.6771417856216431,0.050563108175992966,0.4366490840911865,0.5240100026130676,0.12869438529014587,0.4039788544178009,0.22678259015083313,0.235320582985878,0.2683814465999603,0.26834434270858765,0.2240869551897049,0.17782311141490936,0.08096583932638168,0.4539759159088135,0.13646771013736725,0.5671730041503906,0.22854319214820862,0.5906022191047668,0.03830946236848831,0.2382073849439621,0.19949665665626526,0.253493070602417,0.3172534108161926,0.6353503465652466,0.180657759308815,0.03716561198234558,0.04526771977543831,0.26802653074264526,0.19628934562206268,0.11686427146196365,0.4482332766056061,0.2750421464443207,0.4239201247692108,0.08271268755197525,0.2639942467212677,0.1962183564901352,0.12455058842897415,0.16293112933635712,0.14946788549423218,0.042487192898988724,0.16687892377376556,0.28951945900917053,0.6524286270141602,0.29754772782325745,0.4796634018421173,0.46671149134635925,0.08504541963338852,0.047120291739702225],"values_recorded":64,"members":[{"code_sha256_prefix":"d55eaf6b69be6376","path":"capsule_network.py","papers":["1710.09829"],"paper_pages":[{"arxiv_id":"1710.09829","page":"/paper/dynamic-routing-between-capsules"}],"arg_sig_recorded":[["input",3,"float","float32"]],"scalar_args":{"dim":"1"},"class_bearing":false}]}]},{"bucket":[[4,"float","float32"]],"n_implementations_compared":1,"n_papers":1,"n_not_digested":0,"not_digested_by_type":{},"clusters":[{"output_sha":"e318343bfebd2792","size":1,"n_papers":1,"shape":[2,3,4,4],"dtype":"float64","type":"ndarray","finite":true,"max_difference_among_recorded_values":null,"members_with_recorded_values":1,"torch_reference_conventions_with_this_digest":[],"values":[0.1835346519947052,0.1711222231388092,0.378185510635376,0.2671576142311096,0.44565173983573914,0.11701648682355881,0.1089354082942009,0.32839635014533997,0.3940820097923279,0.2448931187391281,0.08499979227781296,0.27602508664131165,0.14211390912532806,0.18669940531253815,0.020085250958800316,0.6511014699935913,0.16685299575328827,0.29757851362228394,0.47484445571899414,0.060724034905433655,0.056894879788160324,0.3715425133705139,0.0789387971162796,0.49262380599975586,0.22775530815124512,0.5954393744468689,0.023023854941129684,0.1537814885377884,0.2636919915676117,0.18208082020282745,0.4440059959888458,0.11022123694419861,0.14071263372898102,0.5950971841812134,0.1322137862443924,0.1319764107465744,0.05537363141775131,0.7796655297279358,0.13276125490665436,0.03219955042004585,0.1653938740491867,0.04656949266791344,0.09729955345392227,0.6907370686531067,0.287840873003006,0.23885586857795715,0.39240047335624695,0.08090273290872574,0.20672015845775604,0.11852053552865982,0.6024746894836426,0.07228460162878036,0.7849693298339844,0.08038564771413803,0.013502492569386959,0.12114249914884567,0.1549156755208969,0.0833740085363388,0.15777549147605896,0.6039348244667053,0.02647198736667633,0.20719851553440094,0.22983096539974213,0.536498486995697,0.17965355515480042,0.025951389223337173,0.16722643375396729,0.6271686553955078,0.062400199472904205,0.03929351642727852,0.8569360375404358,0.04137016832828522,0.3427213132381439,0.2484167069196701,0.31315985321998596,0.09570213407278061,0.3979133069515228,0.1742614358663559,0.25869905948638916,0.16912616789340973,0.04131356254220009,0.34740832448005676,0.44737720489501953,0.1639009267091751,0.23736397922039032,0.39671608805656433,0.025316735729575157,0.340603232383728,0.12190095335245132,0.15050865709781647,0.6974405646324158,0.030149810016155243,0.4316388666629791,0.3042762875556946,0.06684276461601257,0.1972421407699585],"values_recorded":96,"members":[{"code_sha256_prefix":"442a2012d75c4821","path":"scripts-predict/predict.py","papers":["1901.08394"],"paper_pages":[{"arxiv_id":"1901.08394","page":"/paper/application-of-decision-rules-for-handling"}],"arg_sig_recorded":[["logits",4,"float","float32"]],"scalar_args":{},"class_bearing":false}]}]}]}