{"id":"W4385573830","doi":"10.18653/v1/2022.findings-emnlp.76","title":"Faithful to the Document or to the World? Mitigating Hallucinations via Entity-Linked Knowledge in Abstractive Summarization","year":2022,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Automatic summarization; Computer science; Knowledge base; Information retrieval; Source text; World Wide Web; Data science; Artificial intelligence","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":[0.0008516893,0.00009448524,0.00008528199,0.000181018,0.0005365571,0.0002143779,0.001067874,0.0000132186,0.0002551873],"category_scores_gemma":[0.00008405959,0.00005910084,0.000031873,0.001350351,0.000008611458,0.0002786168,0.0009700949,0.0002283497,0.00008907274],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002570782,"about_ca_system_score_gemma":0.0001002647,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005203438,"about_ca_topic_score_gemma":0.0204883,"domain_scores_codex":[0.9986654,0.0001892318,0.0002648833,0.000341505,0.0003158706,0.0002231163],"domain_scores_gemma":[0.9990264,0.0002769119,0.00005844659,0.0005064503,0.00007462554,0.00005715094],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001837955,0.0002637343,0.0009797637,0.00001352721,0.00003437925,0.00001166158,0.04836667,0.4690083,0.0007676402,0.2851516,0.00652701,0.1888573],"study_design_scores_gemma":[0.0004252137,0.0001085401,0.01583807,0.00003090976,0.000009396797,0.000007827097,0.003336046,0.8787907,0.0009415025,0.005300126,0.09486036,0.0003513235],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02919008,0.00001890055,0.9237925,0.03654423,0.0005758893,0.0008263718,0.000002211484,0.00008328468,0.008966528],"genre_scores_gemma":[0.9647089,5.477853e-7,0.02395279,0.001728704,0.00007948086,0.0003241405,0.00000283324,0.000006811051,0.009195807],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9355188,"threshold_uncertainty_score":0.9973852,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02367705475067786,"score_gpt":0.2922954430037961,"score_spread":0.2686183882531182,"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."}}