{"id":"W4417248049","doi":"10.48550/arxiv.2504.06949","title":"Adaptive Computation Pruning for the Forgetting Transformer","year":2025,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Alliance de recherche numérique du Canada","keywords":"Softmax function; Forgetting; Computation; Pruning; FLOPS; Transformer; Context (archaeology); Set (abstract data type)","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000782351,0.001235203,0.000726355,0.0005480131,0.000436989,0.0008133794,0.002401237,0.0007187729,0.005805325],"category_scores_gemma":[0.003252765,0.0004978919,0.0007651366,0.0005027552,0.0007443691,0.001881853,0.001362321,0.00208241,0.001400966],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009676185,"about_ca_system_score_gemma":0.001739731,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006349972,"about_ca_topic_score_gemma":0.01476868,"domain_scores_codex":[0.9995604,0.00006823531,0.000032354,0.0001230958,0.0001422078,0.00007363719],"domain_scores_gemma":[0.9990793,0.0004243991,0.00005659842,0.0002241187,0.0001612651,0.00005426049],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005642542,0.0001732007,0.002887394,0.0003795657,0.0001712592,0.0003727292,0.0002251553,0.1384195,0.04920723,0.02447685,0.01853005,0.7645928],"study_design_scores_gemma":[0.00005700583,0.00009070764,0.0004552249,0.00002052658,0.00005087645,0.0001540062,0.00002095304,0.9605128,0.02244441,0.01123448,0.004940857,0.00001815331],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02828364,0.0006565116,0.9580376,0.000309623,0.0001850816,0.0001029253,0.0001859031,0.0094784,0.002760321],"genre_scores_gemma":[0.5211095,0.0005122643,0.4673837,0.0006406959,0.00011884,0.0002237263,0.0007017221,0.001036883,0.008272689],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006349972,"threshold_uncertainty_score":0.01942068,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07684360389762872,"score_gpt":0.3255755501200715,"score_spread":0.2487319462224428,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}