{"id":"W4385567355","doi":"10.18653/v1/2022.findings-emnlp.502","title":"Probing for Constituency Structure in Neural Language Models","year":2022,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"Deutsche Forschungsgemeinschaft","keywords":"Treebank; Computer science; Natural language processing; Focus (optics); Artificial intelligence; Syntactic structure; Syntax; Tree structure; Parsing; Data structure; Programming language","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00009494535,0.00005134851,0.00006628354,0.00006079436,0.00008098511,0.00003343715,0.0004271388,0.00001233439,0.00003498945],"category_scores_gemma":[0.000006836216,0.00004842059,0.00002077319,0.000133136,0.000006185861,0.0002371351,0.0002470263,0.0001003081,2.999671e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003820093,"about_ca_system_score_gemma":0.00004359203,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006093012,"about_ca_topic_score_gemma":0.00006209352,"domain_scores_codex":[0.999401,0.00002084367,0.0001081055,0.0002092369,0.0001067478,0.0001540313],"domain_scores_gemma":[0.9997061,0.00002428125,0.00002094718,0.0002184105,0.00001008221,0.00002020027],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000002514525,0.00001263276,0.0001863813,0.00001129224,0.000002016529,0.00001343764,0.003819378,0.4804707,0.001363216,0.4934452,0.00009218031,0.02058113],"study_design_scores_gemma":[0.000196657,0.00001337504,0.000009963978,0.000001316749,4.949071e-7,0.00001126572,0.0001921557,0.9826337,0.0001495436,0.01660345,0.0001232097,0.00006484198],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1503763,0.00009793464,0.8468935,0.0005655743,0.0002347584,0.0002536706,0.000005121484,0.00008851183,0.001484622],"genre_scores_gemma":[0.9212291,1.35766e-7,0.07812307,0.0003870057,0.00001742189,0.00003593987,0.000001810986,0.00000299723,0.0002024658],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7708529,"threshold_uncertainty_score":0.1974534,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02351577902937655,"score_gpt":0.2507793196022692,"score_spread":0.2272635405728926,"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."}}