# Consistency Training Along the Transformer Stack

arXiv: https://arxiv.org/abs/2606.05817 · HTML: https://mlsanity.com/paper/2606.05817/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 43 references that are in the map.
- First posted on arXiv 4 Jun 2026.

## Builds on (most cited first, 25 of 25)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment](https://mlsanity.com/paper/2303.16634/) (2023)
- [Gemma 2: Improving Open Language Models at a Practical Size](https://mlsanity.com/paper/2408.00118/) (2024)
- [Calibrate Before Use: Improving Few-Shot Performance of Language Models](https://mlsanity.com/paper/2102.09690/) (2021)
- [Transformer Feed-Forward Layers Are Key-Value Memories](https://mlsanity.com/paper/2012.14913/) (2020)
- [Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivity](https://mlsanity.com/paper/2104.08786/) (2021)
- [Towards Understanding Sycophancy in Language Models](https://mlsanity.com/paper/2310.13548/) (2023)
- [Enhancing Chat Language Models by Scaling High-quality Instructional Conversations](https://mlsanity.com/paper/2305.14233/) (2023)
- [WildChat: 1M ChatGPT Interaction Logs in the Wild](https://mlsanity.com/paper/2405.01470/) (2024)
- [Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks](https://mlsanity.com/paper/2404.02151/) (2024)
- [Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs](https://mlsanity.com/paper/2407.15549/) (2024)
- [Model Organisms for Emergent Misalignment](https://mlsanity.com/paper/2506.11613/) (2025)
- [Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-time](https://mlsanity.com/paper/2510.04340/) (2025)
- [Inoculation Prompting: Instructing LLMs to misbehave at train-time improves test-time alignment](https://mlsanity.com/paper/2510.05024/) (2025)
- [Bias-Augmented Consistency Training Reduces Biased Reasoning in Chain-of-Thought](https://mlsanity.com/paper/2403.05518/) (2024)
- [Conditional misalignment: common interventions can hide emergent misalignment behind contextual triggers](https://mlsanity.com/paper/2604.25891/) (2026)
- [Consistency Training Helps Stop Sycophancy and Jailbreaks](https://mlsanity.com/paper/2510.27062/) (2025)
- [Gemma Needs Help: Investigating and Mitigating Emotional Instability in LLMs](https://mlsanity.com/paper/2603.10011/) (2026)
- [Consistency Training while Mitigating Obfuscation via Rate Matching](https://mlsanity.com/paper/2606.02211/) (2026)
- [Exposing the Systematic Vulnerability of Open-Weight Models to Prefill Attacks](https://mlsanity.com/paper/2602.14689/) (2026)
- [Consistency Training Can Entrench Misalignment](https://mlsanity.com/paper/2606.03810/) (2026)

## Built on by (most cited first, 1 of 1)

- [How Does Alignment Tuning Shape Representations of Sycophancy and Related Cue-Induced Biases in LLMs?](https://mlsanity.com/paper/2607.18114/) (2026)

## Abstract (arXiv)

Consistency training encourages models to behave similarly across different contexts, and has shown promise for reducing misalignment. We broaden the scope of consistency training in two ways. First, we introduce two new internal consistency targets: MLP Consistency Training (MLPCT), which matches post-activation MLP states, and Attention Consistency Training (AttCT), which matches per-head attention distributions. Second, we apply consistency training to four additional safety threats: persona in-context learning attacks, adversarial frustration, prefill attacks, and conditional misalignment. Across several models and threat settings, we find that consistency training reduces misalignment well beyond the sycophancy and jailbreak settings studied in prior work. We also find cases of cross-threat generalization, where training against one failure mode improves robustness to another, and identify a shared residual-stream mechanism underlying ACT, MLPCT, and AttCT, while distinguishing BCT as mechanistically distinct. Our results suggest that consistency training is a flexible and extensible framework for alignment, capable of unifying defenses against a broader class of model pathologies.