Related papers: Preference Learning with Lie Detectors can Induce …

Activation-based probes have emerged as a promising approach for detecting deceptively aligned AI systems by identifying internal conflict between true and stated goals. We identify a fundamental blind spot: probes fail on coherent…

Direct Preference Optimization (DPO) has emerged as a de-facto approach for aligning language models with human preferences. Recent work has shown DPO's effectiveness relies on training data quality. In particular, clear quality differences…

Recent research on large language models (LLMs) has demonstrated their ability to understand and employ deceptive behavior, even without explicit prompting. However, such behavior has only been observed in rare, specialized cases and has…

Today's high-stakes adversarial interactions feature attackers who constantly breach the ever-improving security measures. Deception mitigates the defender's loss by misleading the attacker to make suboptimal decisions. In order to formally…

Large language models exhibit impressive reasoning capabilities yet frequently generate plausible but incorrect solutions, a phenomenon commonly termed hallucination. This paper investigates the effect of training objective composition on…

Language Models (LMs) may acquire harmful knowledge, and yet feign ignorance of these topics when under audit. Inspired by the recent discovery of deception-related behaviour patterns in LMs, we aim to train classifiers that detect when a…

Large language models (LLMs) can be dishonest when reporting on their actions and beliefs -- for example, they may overstate their confidence in factual claims or cover up evidence of covert actions. Such dishonesty may arise due to the…

Large language models (LLMs) trained on webscale data can produce toxic outputs, raising concerns for safe deployment. Prior defenses, based on applications of DPO, NPO, and similar algorithms, reduce the likelihood of harmful…

In practice, preference learning from human feedback depends on incomplete data with hidden context. Hidden context refers to data that affects the feedback received, but which is not represented in the data used to train a preference…

Aligning large language models (LLMs) with human preferences has gained significant attention, with Proximal Policy Optimization (PPO) as a standard yet computationally expensive method and Direct Preference Optimization (DPO) as a more…

Effective training of language models (LMs) for mathematical reasoning tasks demands high-quality supervised fine-tuning data. Besides obtaining annotations from human experts, a common alternative is sampling from larger and more powerful…

We explore the ability of large language models (LLMs) to engage in subtle deception through strategically phrasing and intentionally manipulating information. This harmful behavior can be hard to detect, unlike blatant lying or…

Deception detection has attracted increasing attention due to its importance in real-world scenarios. Its main goal is to detect deceptive behaviors from multimodal clues such as gestures, facial expressions, prosody, etc. However, these…

Probing has emerged as a promising method for monitoring large language models (LLMs), enabling cheap inference-time detection of concerning behaviours. However, natural examples of many behaviours are rare, forcing researchers to rely on…

LLMs are increasingly used to make or support high-stakes decisions under uncertainty, where alignment depends not only on factual accuracy but on how models weigh tradeoffs between different outcomes. We present an empirical pipeline for…

We propose probe-based data attribution, a method that traces behavioral changes in post-trained language models to responsible training datapoints. By computing activation-difference vectors for both test prompts and preference pairs and…

Large language models (LLMs) exhibit hallucinations (i.e., unfaithful or nonsensical information) when serving as AI assistants in various domains. Since hallucinations always come with truthful content in the LLM responses, previous…

Preference alignment methods such as RLHF and Direct Preference Optimization (DPO) improve instruction following, but they can also reinforce hallucinations when preference judgments reward fluency and confidence over factual correctness.…

Direct Preference Optimization (DPO) and its variants are increasingly used for aligning language models with human preferences. Although these methods are designed to teach a model to generate preferred responses more frequently relative…

Iterative preference optimization methods have recently been shown to perform well for general instruction tuning tasks, but typically make little improvement on reasoning tasks (Yuan et al., 2024, Chen et al., 2024). In this work we…