Beyond Shapley: An Influence-Based Data Auditing Pipeline for LLM Alignment and Evaluation

• URL: http://arxiv.org/abs/2607.22766v1
• Date: Fri, 24 Jul 2026 03:31:44 GMT
• Title: Beyond Shapley: An Influence-Based Data Auditing Pipeline for LLM Alignment and Evaluation
• Abstract summary: We introduce a scalable, inference-only data valuation pipeline that approximates the Shapley value without iterative model retraining.<n>We demonstrate the pipeline's efficacy in sanitizing two heavily vetted alignment datasets.
• Score: 2.733105115068754
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The alignment of Large Language Models (LLMs) is increasingly bottlenecked by data quality. As datasets scale, massive preference and instruction-tuning corpora inevitably accumulate hidden structural contradictions, safety risks, and systemic human annotation errors. Standard dataset auditing methods, such as semantic deduplication or LLM-as-a-judge, struggle to capture the actual predictive impact of individual records and often miss deep functional rule clashes. To address this, we introduce a scalable, inference-only data valuation pipeline that approximates the Shapley value without iterative model retraining. By mapping semantic k-NN neighborhoods into a directed graph, our framework evaluates data utility directly through a reference LLM's probability distribution using zero-shot and one-shot conditional log-likelihood shifts. Our pipeline then translates these predictive influence scores into localized advantage metrics to isolate gradient-conflicting records. We demonstrate the pipeline's efficacy in sanitizing two heavily vetted alignment datasets. First, applying our pipeline to the HelpSteer2 dataset reduced the manual audit search space by 99.1%, successfully uncovering falsely-labeled records across diverse failure modes. Second, applying our automated audit strategy to Anthropic's HH-RLHF training and evaluation splits identified thousands of hidden safety and factual preference inversions. Crucially, by extending this audit to the evaluation split, we expose severe vulnerabilities in current benchmark integrity: highly capable models frequently predict the safer or more helpful response, only to be penalized by objectively flawed human ground-truth labels. Overall, our work provides a mathematically grounded, highly efficient diagnostic tool to uncover human label failures, sanitize evaluation benchmarks, and ensure the integrity of LLM alignment data.

Related papers

• A Step Towards Robust Unsupervised Domain Adaptation via Fine-Tuning and Reinforcement Learning [2.436631469537453]
  We propose textbfSFT+RL, a two-stage robust UDA framework that integratesSupervised Fine Tuning and Reinforcement Learning.<n>textbfSFT+RL achieves average improvements of textbf10.2% in clean accuracy and textbf15.8% in adversarial robustness across three datasets.
  arXiv  Detail & Related papers  (2026-07-03T20:56:12Z)
• CauTion: Knowing When to Trust LLMs for Ensemble Causal Discovery [51.07538881798502]
  Large language models (LLMs) offer a promising source of domain knowledge to complement statistical inference.<n>We propose CauTion, a framework that reliably integrates LLM domain knowledge into an ensemble of statistical causal discovery algorithms.<n>CauTion consistently outperforms both data-centric and LLM-augmented baselines.
  arXiv  Detail & Related papers  (2026-06-02T13:07:43Z)
• Rubrics as an Attack Surface: Stealthy Preference Drift in LLM Judges [39.6038287387348]
  LLM-based judges' behavior is guided by natural rubrics and validated on benchmarks.<n>We identify a previously underrecognized vulnerability in this workflow, which we term-Induced Preference Drift (RIPD)<n>We show this vulnerability can be exploited through preference attacks, in which benchmark-compliant edits steer judgments away from a fixed human or trusted reference on target domains.
  arXiv  Detail & Related papers  (2026-02-14T03:19:14Z)
• Distributionally Robust Optimization with Adversarial Data Contamination [49.89480853499918]
  We focus on optimizing Wasserstein-1 DRO objectives for generalized linear models with convex Lipschitz loss functions.<n>Our primary contribution lies in a novel modeling framework that integrates robustness against training data contamination with robustness against distributional shifts.<n>This work establishes the first rigorous guarantees, supported by efficient computation, for learning under the dual challenges of data contamination and distributional shifts.
  arXiv  Detail & Related papers  (2025-07-14T18:34:10Z)
• Stress-Testing ML Pipelines with Adversarial Data Corruption [11.91482648083998]
  Regulators now demand evidence that high-stakes systems can withstand realistic, interdependent errors.<n>We introduce SAVAGE, a framework that formally models data-quality issues through dependency graphs and flexible corruption templates.<n>Savanage employs a bi-level optimization approach to efficiently identify vulnerable data subpopulations and fine-tune corruption severity.
  arXiv  Detail & Related papers  (2025-06-02T00:41:24Z)
• Scalable Valuation of Human Feedback through Provably Robust Model Alignment [18.151660339831718]
  A robust alignment objective should yield identical model parameters even under severe label noise.<n>We propose H"older-DPO, the first principled alignment loss with a provable redescending property.<n>This metric is gradient-free, enabling scalable and automated human feedback valuation.
  arXiv  Detail & Related papers  (2025-05-23T13:12:37Z)
• Noisy Correspondence Learning with Self-Reinforcing Errors Mitigation [63.180725016463974]
  Cross-modal retrieval relies on well-matched large-scale datasets that are laborious in practice. We introduce a novel noisy correspondence learning framework, namely textbfSelf-textbfReinforcing textbfErrors textbfMitigation (SREM)
  arXiv  Detail & Related papers  (2023-12-27T09:03:43Z)
• Bring Your Own Data! Self-Supervised Evaluation for Large Language Models [52.15056231665816]
  We propose a framework for self-supervised evaluation of Large Language Models (LLMs) We demonstrate self-supervised evaluation strategies for measuring closed-book knowledge, toxicity, and long-range context dependence. We find strong correlations between self-supervised and human-supervised evaluations.
  arXiv  Detail & Related papers  (2023-06-23T17:59:09Z)
• Leveraging Unlabeled Data to Predict Out-of-Distribution Performance [63.740181251997306]
  Real-world machine learning deployments are characterized by mismatches between the source (training) and target (test) distributions. In this work, we investigate methods for predicting the target domain accuracy using only labeled source data and unlabeled target data. We propose Average Thresholded Confidence (ATC), a practical method that learns a threshold on the model's confidence, predicting accuracy as the fraction of unlabeled examples.
  arXiv  Detail & Related papers  (2022-01-11T23:01:12Z)
• WSSOD: A New Pipeline for Weakly- and Semi-Supervised Object Detection [75.80075054706079]
  We propose a weakly- and semi-supervised object detection framework (WSSOD) An agent detector is first trained on a joint dataset and then used to predict pseudo bounding boxes on weakly-annotated images. The proposed framework demonstrates remarkable performance on PASCAL-VOC and MSCOCO benchmark, achieving a high performance comparable to those obtained in fully-supervised settings.
  arXiv  Detail & Related papers  (2021-05-21T11:58:50Z)

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.