phenomenon · first seen 12 Oct 2026
explanation faithfulness
The degree to which a model's stated explanations are supported by the evidence it actually relies on.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Not Every Cue Is Evidence: Cue Supportability for Forensic Distillation in Face Forgery DetectionPapers · 12 Oct 2026 · Binjia Zhou, Dawei Luo, Shuai Chen, Feng Xu, Haoyuan Li, Jia wen Seow, Jiachi Wa…
- Do Language Models Trust Their Own Justifications? A Study on Functional ConsistencyPapers · 23 Oct 2026 · Alisson Rosa Pereira, Bruno Iochins Grisci
- Navigating Disagreement Among Criteria for Evaluating Explanation FaithfulnessPapers · 4 Nov 2026 · Grant Shanklin, Jiří Němeček, Asia Biega
Latest sources
- Navigating Disagreement Among Criteria for Evaluating Explanation FaithfulnessPapers · 4 Nov 2026 · Grant Shanklin, Jiří Němeček, Asia Biega
- Do Language Models Trust Their Own Justifications? A Study on Functional ConsistencyPapers · 23 Oct 2026 · Alisson Rosa Pereira, Bruno Iochins Grisci
- Not Every Cue Is Evidence: Cue Supportability for Forensic Distillation in Face Forgery DetectionPapers · 12 Oct 2026 · Binjia Zhou, Dawei Luo, Shuai Chen, Feng Xu, Haoyuan Li, Jia wen Seow, Jiachi Wa…