Let's Measure Information Step-by-Step: LLM-Based Evaluation Beyond Vibes

• URL: http://arxiv.org/abs/2508.05469v2
• Date: Thu, 21 Aug 2025 17:52:56 GMT
• Title: Let's Measure Information Step-by-Step: LLM-Based Evaluation Beyond Vibes
• Authors: Zachary Robertson, Sanmi Koyejo,
• Abstract summary: We study robustness of AI systems without ground truth by exploiting a link between strategic gaming and information loss.<n>We analyze which information-theoretic mechanisms resist adversarial bounds, extending finite-sample manipulation to show that bounded f-divergences maintain under attacks.
• Score: 14.371259136517802
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We study evaluation of AI systems without ground truth by exploiting a link between strategic gaming and information loss. We analyze which information-theoretic mechanisms resist adversarial manipulation, extending finite-sample bounds to show that bounded f-divergences (e.g., total variation distance) maintain polynomial guarantees under attacks while unbounded measures (e.g., KL divergence) degrade exponentially. To implement these mechanisms, we model the overseer as an agent and characterize incentive-compatible scoring rules as f-mutual information objectives. Under adversarial attacks, TVD-MI maintains effectiveness (area under curve 0.70-0.77) while traditional judge queries are near change (AUC $\approx$ 0.50), demonstrating that querying the same LLM for information relationships rather than quality judgments provides both theoretical and practical robustness. The mechanisms decompose pairwise evaluations into reliable item-level quality scores without ground truth, addressing a key limitation of traditional peer prediction. We release preregistration and code.

Related papers

• The Shadow Self: Intrinsic Value Misalignment in Large Language Model Agents [37.75212140218036]
  We formalize the Loss-of-Control risk and identify the previously underexamined Intrinsic Value Misalignment (Intrinsic VM)<n>We then introduce IMPRESS, a scenario-driven framework for systematically assessing this risk.<n>We evaluate Intrinsic VM on 21 state-of-the-art LLM agents and find that it is a common and broadly observed safety risk across models.
  arXiv  Detail & Related papers  (2026-01-24T07:09:50Z)
• Gaming the Judge: Unfaithful Chain-of-Thought Can Undermine Agent Evaluation [76.5533899503582]
  Large language models (LLMs) are increasingly used as judges to evaluate agent performance.<n>We show this paradigm implicitly assumes that the agent's chain-of-thought (CoT) reasoning faithfully reflects both its internal reasoning and the underlying environment state.<n>We demonstrate that manipulated reasoning alone can inflate false positive rates of state-of-the-art VLM judges by up to 90% across 800 trajectories spanning diverse web tasks.
  arXiv  Detail & Related papers  (2026-01-21T06:07:43Z)
• Explainability-Guided Defense: Attribution-Aware Model Refinement Against Adversarial Data Attacks [6.573058520271728]
  We identify a connection between interpretability and robustness that can be directly leveraged during training.<n>We introduce an attribution-guided refinement framework that transforms Local Interpretable Model-Agnostic Explanations into an active training signal.
  arXiv  Detail & Related papers  (2026-01-02T19:36:03Z)
• DeLeaker: Dynamic Inference-Time Reweighting For Semantic Leakage Mitigation in Text-to-Image Models [55.30555646945055]
  Text-to-Image (T2I) models are vulnerable to semantic leakage.<n>We introduce DeLeaker, a lightweight approach that mitigates leakage by directly intervening on the model's attention maps.<n>SLIM is the first dataset dedicated to semantic leakage.
  arXiv  Detail & Related papers  (2025-10-16T17:39:21Z)
• TrustJudge: Inconsistencies of LLM-as-a-Judge and How to Alleviate Them [58.04324690859212]
  Large Language Models (LLMs) as automated evaluators (LLM-as-a-judge) has revealed critical inconsistencies in current evaluation frameworks.<n>We identify two fundamental types of inconsistencies: Score-Comparison Inconsistency and Pairwise Transitivity Inconsistency.<n>We propose TrustJudge, a probabilistic framework that addresses these limitations through two key innovations.
  arXiv  Detail & Related papers  (2025-09-25T13:04:29Z)
• Incentivizing Truthful Language Models via Peer Elicitation Games [10.530016288072506]
  Large Language Models (LLMs) have demonstrated strong generative capabilities but remain prone to inconsistencies and hallucinations.<n>We introduce Peer Elicitation Games (PEG), a training-free, game-theoretic framework for aligning LLMs through a peer elicitation mechanism involving a generator and multiple discriminators instantiated from distinct base models.
  arXiv  Detail & Related papers  (2025-05-19T18:16:58Z)
• Collaborative Value Function Estimation Under Model Mismatch: A Federated Temporal Difference Analysis [55.13545823385091]
  Federated reinforcement learning (FedRL) enables collaborative learning while preserving data privacy by preventing direct data exchange between agents.<n>In real-world applications, each agent may experience slightly different transition dynamics, leading to inherent model mismatches.<n>We show that even moderate levels of information sharing significantly mitigate environment-specific errors.
  arXiv  Detail & Related papers  (2025-03-21T18:06:28Z)
• Self-Supervised Inference of Agents in Trustless Environments [44.99833362998488]
  We propose a novel approach where agents can form swarms to produce high-quality responses effectively. This is accomplished by utilizing agents capable of data inference and ranking. We show that our approach is an order of magnitude faster than other trustless inference strategies reaching less than 125 ms validation latency.
  arXiv  Detail & Related papers  (2024-09-12T20:32:07Z)
• Towards Effective Evaluations and Comparisons for LLM Unlearning Methods [97.2995389188179]
  This paper seeks to refine the evaluation of machine unlearning for large language models.<n>It addresses two key challenges -- the robustness of evaluation metrics and the trade-offs between competing goals.
  arXiv  Detail & Related papers  (2024-06-13T14:41:00Z)
• In Search of Lost Online Test-time Adaptation: A Survey [40.68806005826287]
  This article presents a comprehensive survey of online test-time adaptation (OTTA) We classify OTTA techniques into three primary categories and benchmark them using a modern backbone, the Vision Transformer (ViT) Our findings diverge from existing literature, revealing that transformers demonstrate heightened resilience to diverse domain shifts.
  arXiv  Detail & Related papers  (2023-10-31T05:47:33Z)
• Distributional Shift-Aware Off-Policy Interval Estimation: A Unified Error Quantification Framework [8.572441599469597]
  We study high-confidence off-policy evaluation in the context of infinite-horizon Markov decision processes. The objective is to establish a confidence interval (CI) for the target policy value using only offline data pre-collected from unknown behavior policies. We show that our algorithm is sample-efficient, error-robust, and provably convergent even in non-linear function approximation settings.
  arXiv  Detail & Related papers  (2023-09-23T06:35:44Z)
• Avoid Adversarial Adaption in Federated Learning by Multi-Metric Investigations [55.2480439325792]
  Federated Learning (FL) facilitates decentralized machine learning model training, preserving data privacy, lowering communication costs, and boosting model performance through diversified data sources. FL faces vulnerabilities such as poisoning attacks, undermining model integrity with both untargeted performance degradation and targeted backdoor attacks. We define a new notion of strong adaptive adversaries, capable of adapting to multiple objectives simultaneously. MESAS is the first defense robust against strong adaptive adversaries, effective in real-world data scenarios, with an average overhead of just 24.37 seconds.
  arXiv  Detail & Related papers  (2023-06-06T11:44:42Z)
• Divide, Conquer, and Combine: Mixture of Semantic-Independent Experts for Zero-Shot Dialogue State Tracking [83.40120598637665]
  Zero-shot transfer learning for Dialogue State Tracking (DST) helps to handle a variety of task-oriented dialogue domains without the cost of collecting in-domain data. Existing works mainly study common data- or model-level augmentation methods to enhance the generalization. We present a simple and effective "divide, conquer and combine" solution, which explicitly disentangles the semantics of seen data.
  arXiv  Detail & Related papers  (2023-06-01T08:21:20Z)
• Byzantine-Robust Online and Offline Distributed Reinforcement Learning [60.970950468309056]
  We consider a distributed reinforcement learning setting where multiple agents explore the environment and communicate their experiences through a central server. $alpha$-fraction of agents are adversarial and can report arbitrary fake information. We seek to identify a near-optimal policy for the underlying Markov decision process in the presence of these adversarial agents.
  arXiv  Detail & Related papers  (2022-06-01T00:44:53Z)
• Efficient Model-based Multi-agent Reinforcement Learning via Optimistic Equilibrium Computation [93.52573037053449]
  H-MARL (Hallucinated Multi-Agent Reinforcement Learning) learns successful equilibrium policies after a few interactions with the environment. We demonstrate our approach experimentally on an autonomous driving simulation benchmark.
  arXiv  Detail & Related papers  (2022-03-14T17:24:03Z)
• Controllable Guarantees for Fair Outcomes via Contrastive Information Estimation [32.37031528767224]
  Controlling bias in training datasets is vital for ensuring equal treatment, or parity, between different groups in downstream applications. We demonstrate an effective method for controlling parity through mutual information based on contrastive information estimators. We test our approach on UCI Adult and Heritage Health datasets and demonstrate that our approach provides more informative representations across a range of desired parity thresholds.
  arXiv  Detail & Related papers  (2021-01-11T18:57:33Z)
• Trust but Verify: Assigning Prediction Credibility by Counterfactual Constrained Learning [123.3472310767721]
  Prediction credibility measures are fundamental in statistics and machine learning. These measures should account for the wide variety of models used in practice. The framework developed in this work expresses the credibility as a risk-fit trade-off.
  arXiv  Detail & Related papers  (2020-11-24T19:52:38Z)
• Provably Efficient Causal Reinforcement Learning with Confounded Observational Data [135.64775986546505]
  We study how to incorporate the dataset (observational data) collected offline, which is often abundantly available in practice, to improve the sample efficiency in the online setting. We propose the deconfounded optimistic value iteration (DOVI) algorithm, which incorporates the confounded observational data in a provably efficient manner.
  arXiv  Detail & Related papers  (2020-06-22T14:49:33Z)
• Improving LIME Robustness with Smarter Locality Sampling [0.0]
  We propose to make LIME more robust by training a generative adversarial network to sample more realistic synthetic data. Our experiments demonstrate an increase in accuracy across three real-world datasets in detecting biased, adversarial behavior. This is achieved while maintaining comparable explanation quality, with up to 99.94% in top-1 accuracy in some cases.
  arXiv  Detail & Related papers  (2020-06-22T14:36:08Z)

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.