AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Self-explaining deep models with logic rule reasoning

Seungeon Lee, Xiting Wang, Sungwon Han, Xiaoyuan Yi, Xing Xie, Meeyoung Cha

NeurIPS 2022 · Large language models › Language, safety and interpretability · approach: Faithfulness of LLM explanations · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Replace direct backbone prediction with SELOR logic-rule reasoningBERT · YelpAUC: 97.39 -> 97.28No clear effect
Replace direct backbone prediction with SELOR logic-rule reasoningRoBERTa · ClickbaitAUC: 63.72 -> 64.14Improved
Replace direct backbone prediction with SELOR logic-rule reasoningDNN · AdultAUC: 68.62 -> 70.36Improved
Use SELOR instead of RCN with a predefined rule setClickbait · ClickbaitAUC: 59.91 -> 60.31 (BERT); 59.37 -> 64.14 (RoBERTa)Improved
Use SELOR logic-rule explanations instead of LIME explanationsYelp · Yelpgood explanations (%): 89.8 -> 94.4Improved
Use SELOR logic-rule explanations instead of RCN explanationsAdult · Adultbest explanations (%): 10.2 -> 65.1Improved
Use the neural consequent estimator to assess rule confidence across the training dataSELOR · Noisy-label experimentsprediction performance: 2.4% to 13.7%Improved
Exclude the candidate atom vegas during testing (ablation)SELOR · 169 Yelp samples that previously used vegasprediction performanceMixed
Use predefined rules rather than SELOR's generated rulesRCN · ClickbaitAUC: 60.31 vs 59.91 (BERT); 64.14 vs 59.37 (RoBERTa)Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy