{"id":"W7018279385","doi":"","title":"The Cure or the Curse: Investigating the Role and Challenges of Positional Encoding in Transformers","year":2024,"lang":"en","type":"dissertation","venue":"eScholarship@McGill (McGill)","topic":"Categorization, perception, and language","field":"Psychology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Encoding (memory); Transformer; Pattern recognition (psychology); Key (lock)","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003237678,0.0009728191,0.0009892503,0.0008247249,0.0007549868,0.003530152,0.001653314,0.00129904,0.00472255],"category_scores_gemma":[0.02718828,0.0006824583,0.001162346,0.001150611,0.002987992,0.01296161,0.004088252,0.003900917,0.001943832],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001420903,"about_ca_system_score_gemma":0.00274932,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00422285,"about_ca_topic_score_gemma":0.005844669,"domain_scores_codex":[0.9979036,0.0009395382,0.0001578401,0.0004500553,0.0003883585,0.0001606113],"domain_scores_gemma":[0.9894164,0.006625614,0.0005351678,0.002345623,0.0008090485,0.0002681935],"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":[0.0005335776,0.0001542909,0.007505021,0.0006249474,0.0001367212,0.0004820123,0.002527355,0.09653046,0.01081003,0.4253294,0.01057665,0.4447895],"study_design_scores_gemma":[0.00004411599,0.0002153743,0.00066632,0.0001276523,0.00008213508,0.0004046221,0.0006248827,0.491245,0.01096547,0.4833958,0.012172,0.00005654183],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.09543936,0.001205262,0.8892543,0.003359203,0.0002195651,0.00008250959,0.0006178301,0.003164041,0.006657863],"genre_scores_gemma":[0.7059327,0.001604893,0.2814505,0.001009692,0.00015492,0.0001245958,0.001416476,0.001436076,0.006870173],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00472255,"threshold_uncertainty_score":0.01712269,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02549892781871503,"score_gpt":0.2878264370778905,"score_spread":0.2623275092591755,"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."}}