arXiv AI September 30, 2026 By Usman Anwar, Sahar Abdelnabi, David Krueger

Training LLMs to Verbalize Evaluation Awareness

The paper introduces Verbalization Training (VT), a technique that encourages large language models (LLMs) to openly express their evaluation awareness (EA) without directly supervising their internal beliefs. VT works by truncating model rollouts just before spontaneous verbalizations, creating training prefixes that signal awareness, and then applying a reinforcement learning objective to increase calibrated verbalization. Experiments on models such as Qwen3.6-35B-A3B, Kimi K2.6, and Inkling show that VT boosts verbalized EA by 2.4–2.9× while keeping latent EA and overall behavior largely unchanged, and a causal study confirms that VT-induced verbalizations reflect newly acquired meta‑knowledge.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

arXiv Machine Learning

Sep 22

Evaluation Awareness Shifts from Format to Context with Model Scale

arXiv:2609.22119v1 Announce Type: cross Abstract: Evaluation awareness poses an unprecedented threat to model evaluation, but the mechanisms by which models detect it remain unknown. This study focus...

By Navraj Singh, Maheep Chaudhary

llms

arXiv AI

Aug 25

Evaluation Awareness in Language Models: Representation, Verbalization, and Control

arXiv:2608.21766v1 Announce Type: cross Abstract: Both capability and safety benchmarks rest upon the assumption that the behavior of language models undergoing a test is informative about their beha...

By Farzaneh Heidari, Amin Memarian, Guillaume Rabusseau

llms fine-tuning benchmarks safety

arXiv AI

Jul 7

The Rise of Verbal Tics in Large Language Models: A Systematic Analysis Across Frontier Models

arXiv:2604. 19139v3 Announce Type: replace-cross Abstract: As Large Language Models (LLMs) continue to evolve through alignment techniques such as Reinforcement Learning from Human Feedback (RLHF) and Constitutional AI, a growing and increasingly conspicuous phenomenon has emerged: the proliferation of verbal tics--repetitive, formulaic linguistic patterns that pervade model outputs.

By Shuai Wu, Xue Li, Yanna Feng, Yufang Li, Zhijun Wang, Ran Wang

llms reinforcement-learning benchmarks safety

arXiv AI

Aug 12

Evaluation-Conditioned Training: Teaching Models to Generalize to Stronger Oversight Regimes

arXiv:2608. 10209v1 Announce Type: new Abstract: Feedback signals used to train Large Language Models (LLMs) are the primary driver of their behavior and our main lever for instilling alignment with human values and objectives.

By Alec Harris, Kasey Corra, Archie Chaudhury, Yixiong Hao

llms reinforcement-learning fine-tuning safety

arXiv AI

Jun 11

Models That Know How Evaluations Are Designed Score Safer

arXiv:2605. 28591v2 Announce Type: replace-cross Abstract: The validity of AI safety evaluations depends on models behaving consistently across controlled and deployment settings.

By Katharina Deckenbach, Haritz Puerto, Jonas Geiping, Sahar Abdelnabi

fine-tuning benchmarks safety

arXiv AI

Aug 25

Measuring Activation Control in Large Language Models

arXiv:2608.21664v1 Announce Type: new Abstract: Safe deployment of increasingly capable models will likely come to rely on latent-space monitoring as a complement to behavioral evaluations, especiall...

By Marek Mateusz Kowalski, Joshua Fonseca Rivera, Uzay Macar, David Demitri Africa

llms diffusion benchmarks

One email a morning, machine-written

Related stories

Evaluation Awareness Shifts from Format to Context with Model Scale

Evaluation Awareness in Language Models: Representation, Verbalization, and Control

The Rise of Verbal Tics in Large Language Models: A Systematic Analysis Across Frontier Models

Evaluation-Conditioned Training: Teaching Models to Generalize to Stronger Oversight Regimes

Models That Know How Evaluations Are Designed Score Safer

Measuring Activation Control in Large Language Models