{"id":"W4389095096","doi":"10.26434/chemrxiv-2023-r1qn9","title":"Diagram-based Input for Large Language Models to Support Accessible STEM Learning","year":2023,"lang":"en","type":"preprint","venue":"ChemRxiv","topic":"Digital Accessibility for Disabilities","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dow Chemical (Canada)","funders":"National Science Foundation","keywords":"Interfacing; Computer science; Usability; Diagram; Human–computer interaction; Programming language; Database","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.002040922,0.0003758896,0.0005788312,0.0002088781,0.0004614367,0.001005287,0.001382498,0.0004565452,0.0001871635],"category_scores_gemma":[0.0009291862,0.0003948418,0.000459481,0.0004600657,0.0001888528,0.0005656665,0.001037919,0.000506047,0.0002151605],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004078135,"about_ca_system_score_gemma":0.0009385871,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001194535,"about_ca_topic_score_gemma":0.003547436,"domain_scores_codex":[0.9966385,0.0001189235,0.0005418939,0.001025875,0.0006923075,0.0009825436],"domain_scores_gemma":[0.9978635,0.0005528071,0.0002380104,0.0007178236,0.000259282,0.0003685924],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0006609521,0.003374773,0.151755,0.01651619,0.0008458222,0.0000970421,0.4240162,0.06336626,0.0004303067,0.1484989,0.07311834,0.1173202],"study_design_scores_gemma":[0.004968346,0.00102497,0.003563018,0.002868879,0.0005974188,7.648974e-7,0.2522321,0.04000653,0.02130488,0.4876981,0.1767544,0.008980561],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.921266,0.0001096848,0.009584773,0.001992597,0.001414291,0.002965646,0.0002729209,0.001871824,0.06052223],"genre_scores_gemma":[0.9669309,0.00000892407,0.0004866879,0.0002824542,0.0004310863,0.001111451,0.00035989,0.0001030172,0.03028563],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3391992,"threshold_uncertainty_score":0.9998503,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1170541567878942,"score_gpt":0.377402435144251,"score_spread":0.2603482783563568,"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."}}