# Consistency Training while Mitigating Obfuscation via Rate Matching

arXiv: https://arxiv.org/abs/2606.02211 · HTML: https://mlsanity.com/paper/2606.02211/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 25 references that are in the map.
- First posted on arXiv 1 Jun 2026.

## Builds on (most cited first, 16 of 16)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [gpt-oss-120b & gpt-oss-20b Model Card](https://mlsanity.com/paper/2508.10925/) (2025)
- [Discovering Language Model Behaviors with Model-Written Evaluations](https://mlsanity.com/paper/2212.09251/) (2022)
- [Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation](https://mlsanity.com/paper/2503.11926/) (2025)
- [Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety](https://mlsanity.com/paper/2507.11473/) (2025)
- [Analyzing the Generalization and Reliability of Steering Vectors](https://mlsanity.com/paper/2407.12404/) (2024)
- [Large Language Models Often Know When They Are Being Evaluated](https://mlsanity.com/paper/2505.23836/) (2025)
- [Stress Testing Deliberative Alignment for Anti-Scheming Training](https://mlsanity.com/paper/2509.15541/) (2025)
- [Designing a Dashboard for Transparency and Control of Conversational AI](https://mlsanity.com/paper/2406.07882/) (2024)
- [Bias-Augmented Consistency Training Reduces Biased Reasoning in Chain-of-Thought](https://mlsanity.com/paper/2403.05518/) (2024)
- [Probing and Steering Evaluation Awareness of Language Models](https://mlsanity.com/paper/2507.01786/) (2025)
- [Steering Evaluation-Aware Language Models to Act Like They Are Deployed](https://mlsanity.com/paper/2510.20487/) (2025)
- [Consistency Training Helps Stop Sycophancy and Jailbreaks](https://mlsanity.com/paper/2510.27062/) (2025)
- [Output Supervision Can Obfuscate the Chain of Thought](https://mlsanity.com/paper/2511.11584/) (2025)

## Built on by (most cited first, 4 of 4)

- [Consistency Training Can Entrench Misalignment](https://mlsanity.com/paper/2606.03810/) (2026)
- [Consistency Training Along the Transformer Stack](https://mlsanity.com/paper/2606.05817/) (2026)
- [Position: It's Time to Optimize LLMs for Self-Consistency](https://mlsanity.com/paper/2608.05188/) (2026)
- [Unlearning Deceptive Behaviors in LLMs with Contrastive Forget Sets](https://mlsanity.com/paper/2609.38909/) (2026)

## Abstract (arXiv)

Large language models are often influenced by extraneous input features, such as cues revealing a user's preferred answer. Consistency training reduces this influence by training models to behave similarly across inputs with and without the extraneous feature. However, existing methods train for consistency over entire responses or internal activations, which also constrains whether the model verbalises said extraneous features. We show this leads to obfuscation, where the model learns not to mention a cue while remaining influenced by it, which may undermine monitorability. To address this, we introduce Rate Matching Consistency Training (RMCT), which trains for consistency over selected behavioural properties without constraining how this behaviour is expressed. RMCT matches the rate at which the model exhibits a target behaviour (e.g., following a bias cue) across input perturbations, rather than requiring paired inputs with and without the extraneous feature, extending consistency training to settings where the extraneous features cannot be removed. We evaluate RMCT on sycophancy reduction in two open-weight language models, achieving reductions in bias-following comparable to a standard consistency-training baseline on held-out bias types, while largely preserving the model's tendency to verbalise the bias cue. Further, we find that RMCT is more data-efficient at the expense of being less compute-efficient in our experiments. Overall, RMCT shows that consistency training can improve behavioural robustness without directly trading off against monitorability.