From Reweighting to Rewriting: Unlocking the Intervention Effects of Influential Samples in Training Data Attribution

• URL: http://arxiv.org/abs/2609.02771v1
• Date: Wed, 02 Sep 2026 16:07:21 GMT
• Title: From Reweighting to Rewriting: Unlocking the Intervention Effects of Influential Samples in Training Data Attribution
• Abstract summary: Training data attribution aims to identify training examples that shape model behavior.<n>Its intervention value depends on both which examples are selected and how they are modified.
• Score: 31.17710939532687
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Training data attribution (TDA) aims to identify training examples that shape model behavior, but its intervention value depends on both which examples are selected and how they are modified. Influence functions (IF) estimate behavioral changes under infinitesimal reweighting, yet IF-selected examples often show limited advantages over random selection under conventional weight-based interventions. This raises the question of whether influential examples lack intervention value or whether reweighting fails to realize their behavioral leverage.We introduce influence-guided response rewriting, which uses IF to identify intervention targets and replaces their responses with behavior-aligned or behavior-opposed supervision while keeping instructions fixed. Across four open-weight LLMs, we compare rewriting and reweighting on the same influence-selected examples using epistemic abstention as our primary testbed. Response rewriting produces stronger, more persistent, and bidirectional behavioral shifts, while reweighting the same examples yields weak and inconsistent effects. Further analyses show that influence-selected examples provide greater rewriting leverage than alternative selectors, with changes remaining concentrated on target-relevant behaviors. The same qualitative contrast extends to safety refusal. These results distinguish the local reweighting effects captured by influence estimates from the broader intervention leverage of the examples they identify, motivating intervention-aware evaluation of TDA methods.

Related papers

• The Illusion of Intervention: Your LLM-Simulated Experiment is an Observational Study [56.649987197786096]
  Large language models (LLMs) show potential as simulators of human behavior.<n>LLMs can induce unintended shifts in latent user attributes, causing user drift.<n>We formalize the confounding or selection bias that can arise due to user drift.
  arXiv  Detail & Related papers  (2026-05-20T06:09:41Z)
• Attention Hijacking: Response Manipulation Across Queries in Vision-Language Models [57.870323273127234]
  Motivated by the observation, we propose textbfAttention Hijacking, a novel adversarial attack that steers internal attention toward a persistent image-dominant pattern.<n>Our method reduces the dependence of the manipulated output on the specific wording of the query.
  arXiv  Detail & Related papers  (2026-05-17T08:02:27Z)
• Interaction-Aware Influence Functions for Group Attribution [26.126999503061345]
  Influence functions approximate how removing a training example changes a quantity of interest, called the target function, such as a held-out loss.<n>We propose an interaction-aware influence function that characterizes how interactions between examples influence the target.
  arXiv  Detail & Related papers  (2026-05-15T06:54:15Z)
• ConfoundingSHAP: Quantifying confounding strength in causal inference [70.18600230387723]
  In causal inference, confounders are variables that influence both treatment decisions and outcomes.<n>We introduce ConfoundingSHAP, a Shapley-based method for attributing confounding strength to individual covariates.<n>We demonstrate the practical value across various datasets.
  arXiv  Detail & Related papers  (2026-05-11T13:19:33Z)
• Unifying Perplexing Behaviors in Modified BP Attributions through Alignment Perspective [61.5509267439999]
  We present a unified theoretical framework for methods like GBP, RectGrad, LRP, and DTD.<n>We demonstrate that they achieve input alignment by combining the weights of activated neurons.<n>This alignment improves the visualization quality and reduces sensitivity to weight randomization.
  arXiv  Detail & Related papers  (2025-03-14T07:58:26Z)
• Indiscriminate Disruption of Conditional Inference on Multivariate Gaussians [60.22542847840578]
  Despite advances in adversarial machine learning, inference for Gaussian models in the presence of an adversary is notably understudied. We consider a self-interested attacker who wishes to disrupt a decisionmaker's conditional inference and subsequent actions by corrupting a set of evidentiary variables. To avoid detection, the attacker also desires the attack to appear plausible wherein plausibility is determined by the density of the corrupted evidence.
  arXiv  Detail & Related papers  (2024-11-21T17:46:55Z)
• Explaining Knock-on Effects of Bias Mitigation [13.46387356280467]
  In machine learning systems, bias mitigation approaches aim to make outcomes fairer across privileged and unprivileged groups. In this paper, we aim to characterise impacted cohorts when mitigation interventions are applied. We examine a range of bias mitigation strategies that work at various stages of the model life cycle. We show that all tested mitigation strategies negatively impact a non-trivial fraction of cases, i.e., people who receive unfavourable outcomes solely on account of mitigation efforts.
  arXiv  Detail & Related papers  (2023-12-01T18:40:37Z)
• Causal Inference from Text: Unveiling Interactions between Variables [20.677407402398405]
  Existing methods only account for confounding covariables that affect both treatment and outcome. This bias arises from insufficient consideration of non-confounding covariables. In this work, we aim to mitigate the bias by unveiling interactions between different variables.
  arXiv  Detail & Related papers  (2023-11-09T11:29:44Z)
• Causal Ordering Without Effect Estimation: A Framework for Using Proxies in Treatment Prioritization [3.0509197593879844]
  We develop a decision-focused framework to reason about predictive proxies.<n>We identify conditions under which proxies recover the correct effect ordering, which hold when a proxy reflects a dominant moderator of treatment effects.<n>We show how these conditions emerge as a useful approximation in discrete choice settings, where the propensity to act without an intervention moderates persuasion.
  arXiv  Detail & Related papers  (2022-06-25T02:15:22Z)
• Revisiting Methods for Finding Influential Examples [2.094022863940315]
  Methods for finding influential training examples for test-time decisions have been proposed. In this paper, we show that all of the above methods are unstable. We propose to evaluate such explanations by their ability to detect poisoning attacks.
  arXiv  Detail & Related papers  (2021-11-08T18:00:06Z)
• Deconfounding Scores: Feature Representations for Causal Effect Estimation with Weak Overlap [140.98628848491146]
  We introduce deconfounding scores, which induce better overlap without biasing the target of estimation. We show that deconfounding scores satisfy a zero-covariance condition that is identifiable in observed data. In particular, we show that this technique could be an attractive alternative to standard regularizations.
  arXiv  Detail & Related papers  (2021-04-12T18:50:11Z)
• RelatIF: Identifying Explanatory Training Examples via Relative Influence [13.87851325824883]
  We use influence functions to identify relevant training examples that one might hope "explain" the predictions of a machine learning model. We introduce RelatIF, a new class of criteria for choosing relevant training examples by way of an optimization objective that places a constraint on global influence. In empirical evaluations, we find that the examples returned by RelatIF are more intuitive when compared to those found using influence functions.
  arXiv  Detail & Related papers  (2020-03-25T20:59:54Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.