Newest papers

From Outcomes to Components: A Counterfactual Evaluation Framework for LLM-Proposed Repairs

Hengyu Ye, Songhua Hu · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26650 · cited by 0 works (OpenAlex) · open-access PDF

French

Une action de correction de taux de cible sélectionnée par un grand modèle de langue (LLM) spécifie quels enregistrements doivent être modifiés et comment modifier leurs valeurs. Une action échouée peut toujours contenir un choix utile des enregistrements cibles ou une proposition de mise à jour utile, en fonction de la façon dont les deux sont associés. Nous introduisons un cadre d'évaluation contrefaçon qui sépare les enregistrements sélectionnés (objectifs) du changement appliqué (une mise à jour). Quatre exécutions combinent les objectifs proposés et les mises à jour de référence pour mesurer les effets de la correction ou de la rétention de chaque composant et de leur interaction. Nous évaluons le cadre sur deux tâches: attribution de trafic, où la demande affectée doit

Machine translation (Meta M2M100); the English original follows.

Abstract

A repair action proposed by a large language model (LLM) specifies which records to change and how to change their values. A failed action may still contain a useful choice of records or a useful update, depending on how the two are paired. We introduce a counterfactual evaluation framework that separates selected records (targets) from the applied change (an update). Four executions combine proposed and reference targets and updates to measure the effects of correcting or retaining each component and their interaction. We evaluate the framework on two tasks: traffic assignment, where affected demand must be identified from road counts, and database repair, where affected rows must be identified from aggregate reports. An injected fault supplies a reference action that reverses it, and the recovery rate measures an action's loss reduction relative to that reference. For sparse-target actions in the traffic experiments, retaining the proposed update on reference targets achieves 79.4% of the reference loss reduction, compared with 43.7% from tuning update strength on the proposed targets. In the database experiments, update replacement worsens 17 of 108 actions, all selecting unaffected records, with sharply different frequencies across models. The value of a proposed update depends on the targets it is paired with. Missed targets hide useful updates, while excess targets can make a milder update better than the reference correction. In a component-reuse study, testing new target-update pairings from two proposals raises mean held-out recovery rate to 0.616, compared with 0.580 when selecting between the original proposals.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Counterfactual thinking, Set (abstract data type), Aggregate (composite), Action (physics), Computer science, Aggregate data

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange