Subliminal Signals in Preference Labels

• URL: http://arxiv.org/abs/2603.01204v1
• Date: Sun, 01 Mar 2026 17:55:57 GMT
• Title: Subliminal Signals in Preference Labels
• Abstract summary: We show that preference labels can function as a covert communication channel.<n>We show that even when a neutral student model generates semantically unbiased completions, a biased judge can transmit unintended behavioral traits through preference assignments.
• Score: 29.981556395184285
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: As AI systems approach superhuman capabilities, scalable oversight increasingly relies on LLM-as-a-judge frameworks where models evaluate and guide each other's training. A core assumption is that binary preference labels provide only semantic supervision about response quality. We challenge this assumption by demonstrating that preference labels can function as a covert communication channel. We show that even when a neutral student model generates semantically unbiased completions, a biased judge can transmit unintended behavioral traits through preference assignments, which even strengthen across iterative alignment rounds. Our findings suggest that robust oversight in superalignment settings requires mechanisms that can detect and mitigate subliminal preference transmission, particularly when judges may pursue unintended objectives.

Related papers

• Bias Leaves a Gradient Trail: Label-Free Bias Identification via Gradient Probes on Concept Decompositions [8.894428198928333]
  We present a bias-label-free, post-hoc method for identifying spurious concepts in frozen vision models.<n>On Colored MNIST and Waterbirds the method recovers concepts aligned with the known spurious cue.<n> suppressing the top-ranked concepts at inference time improves worst-group accuracy by up to 17.9 percentage points.
  arXiv  Detail & Related papers  (2026-05-27T17:39:10Z)
• MBD: A Model-Based Debiasing Framework Across User, Content, and Model Dimensions [50.00784452900918]
  We propose a general, model-based debiasing (MBD) framework that addresses this challenge.<n>We explicitly estimate the contextual mean and variance of the engagement distribution for arbitrary cohorts.<n>This integration allows the framework to convert biased raw signals into unbiased representations.
  arXiv  Detail & Related papers  (2026-03-15T15:07:01Z)
• Optimizing In-Context Demonstrations for LLM-based Automated Grading [31.353360036776976]
  GUIDE (Grading Using Iteratively Designed Exemplars) is a framework that reframes exemplar selection and refinement as a boundary-focused optimization problem.<n>We show that GUIDE significantly outperforms standard retrieval baselines in experiments in physics, chemistry, and pedagogical content knowledge.
  arXiv  Detail & Related papers  (2026-02-28T04:52:38Z)
• 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)
• Making Bias Non-Predictive: Training Robust LLM Judges via Reinforcement Learning [91.8584139564909]
  Large language models (LLMs) increasingly serve as automated judges, yet they remain susceptible to cognitive biases.<n>We propose Epistemic Independence Training (EIT), a reinforcement learning framework grounded in a key principle.<n>EIT operationalizes this through a balanced conflict strategy where bias signals are equally likely to support correct and incorrect answers.
  arXiv  Detail & Related papers  (2026-02-02T01:43:48Z)
• 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)
• Judging with Confidence: Calibrating Autoraters to Preference Distributions [56.17041629492863]
  We argue that a reliable autorater must learn to model the full distribution of preferences defined by a target population.<n>We present two learning methods tailored to different data conditions.<n>Our results show that finetuning autoraters with a distribution-matching objective leads to verbalized probability predictions that are better aligned with the target preference distribution.
  arXiv  Detail & Related papers  (2025-09-30T20:36:41Z)
• Reference-Free Rating of LLM Responses via Latent Information [53.463883683503106]
  We study the common practice of asking a judge model to assign Likert-scale scores to free-text responses.<n>We then propose and evaluate Latent Judges, which derive scalar ratings from internal model signals.<n>Across a broad suite of pairwise and single-rating benchmarks, latent methods match or surpass standard prompting.
  arXiv  Detail & Related papers  (2025-09-29T12:15:52Z)
• When Does Confidence-Based Cascade Deferral Suffice? [69.28314307469381]
  Cascades are a classical strategy to enable inference cost to vary adaptively across samples. A deferral rule determines whether to invoke the next classifier in the sequence, or to terminate prediction. Despite being oblivious to the structure of the cascade, confidence-based deferral often works remarkably well in practice.
  arXiv  Detail & Related papers  (2023-07-06T04:13:57Z)
• Self-supervised debiasing using low rank regularization [59.84695042540525]
  Spurious correlations can cause strong biases in deep neural networks, impairing generalization ability. We propose a self-supervised debiasing framework potentially compatible with unlabeled samples. Remarkably, the proposed debiasing framework significantly improves the generalization performance of self-supervised learning baselines.
  arXiv  Detail & Related papers  (2022-10-11T08:26:19Z)

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.