Metraipolis › archive › dc_64a615c46e7e

An Empirical Study of Counterfactual Self-Explanations in LLMs — arXiv:2609.17119

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.17119
License (as archived)
CC0-1.0
Archived
2026-10-15T19:19:25.022Z
SHA-256 of the archived text
903f32215289be7609d0cee6d0083034ff3c48a7435b202776c36fbf5bea50ef
Stable URL of this page
https://metraipolis.com/documento/dc_64a615c46e7e

How to cite

An Empirical Study of Counterfactual Self-Explanations in LLMs — arXiv:2609.17119 — arXiv (en), https://arxiv.org/abs/2609.17119 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_64a615c46e7e (Metraipolis archive dc_64a615c46e7e, sha256 903f32215289be7609d0cee6d0083034ff3c48a7435b202776c36fbf5bea50ef)
BibTeX
@misc{dc64a615c46e7e,
  title = {An Empirical Study of Counterfactual Self-Explanations in LLMs — arXiv:2609.17119},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.17119},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_64a615c46e7e; sha256 903f32215289be7609d0cee6d0083034ff3c48a7435b202776c36fbf5bea50ef},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.17119. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models can easily generate explanations for their own outputs, but such self-explanations are not necessarily faithful to the model's behavior. We study this issue through counterfactual self-explanations, where a model minimally edits an input so that its own prediction changes. Across sentiment analysis and natural language inference, we evaluate ten instruction-tuned models from the LLaMA-3 and Qwen-2.5 families, measuring faithfulness, minimality, and alignment with human-annotated rationales. Our results show that model scale is the strongest determinant of explanation quality: larger models are substantially more likely to generate counterfactuals that flip their own predictions and target decision-relevant evidence. In contrast, the rationale-guided condition produces edit-minimal counterfactuals that are also more human-aligned. However, it does not consistently improve faithfulness. Overall, counterfactual self-explanations can provide useful behavioral evidence about model decisions, but their reliability depends strongly on model capacity and should be empirically validated rather than assumed.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-16 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.17119 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T18:32:32.856Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-16?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).