# Rethinking harmless refusals when fine-tuning foundation models

arXiv: https://arxiv.org/abs/2406.19552 · HTML: https://mlsanity.com/paper/2406.19552/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning.
- Builds on 15 of its 22 references that are in the map.
- First posted on arXiv 27 Jun 2024.

## Placed in

- LLM reasoning / [Prompting and chain-of-thought reasoning](https://mlsanity.com/lr/prompting-and-chain-of-thought-reasoning/)

## Builds on (most cited first, 15 of 15)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [Constitutional AI: Harmlessness from AI Feedback](https://mlsanity.com/paper/2212.08073/) (2022)
- [Jailbroken: How Does LLM Safety Training Fail?](https://mlsanity.com/paper/2307.02483/) (2023)
- [Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting](https://mlsanity.com/paper/2305.04388/) (2023)
- [A General Language Assistant as a Laboratory for Alignment](https://mlsanity.com/paper/2112.00861/) (2021)
- [Red Teaming Language Models with Language Models](https://mlsanity.com/paper/2202.03286/) (2022)
- [Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned](https://mlsanity.com/paper/2209.07858/) (2022)
- [Improving alignment of dialogue agents via targeted human judgements](https://mlsanity.com/paper/2209.14375/) (2022)
- [Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training](https://mlsanity.com/paper/2401.05566/) (2024)
- [Scalable and Transferable Black-Box Jailbreaks for Language Models via Persona Modulation](https://mlsanity.com/paper/2311.03348/) (2023)
- [Large Language Models can Strategically Deceive their Users when Put Under Pressure](https://mlsanity.com/paper/2311.07590/) (2023)

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

- [Whose Name Comes Up? II: Benchmarking and Intervention-Based Auditing of LLM-Based Scholar Recommendation](https://mlsanity.com/paper/2602.08873/) (2026)

## Abstract (arXiv)

In this paper, we investigate the degree to which fine-tuning in Large Language Models (LLMs) effectively mitigates versus merely conceals undesirable behavior. Through the lens of semi-realistic role-playing exercises designed to elicit such behaviors, we explore the response dynamics of LLMs post fine-tuning interventions. Our methodology involves prompting models for Chain-of-Thought (CoT) reasoning and analyzing the coherence between the reasoning traces and the resultant outputs. Notably, we identify a pervasive phenomenon we term reason-based deception, where models either stop producing reasoning traces or produce seemingly ethical reasoning traces that belie the unethical nature of their final outputs. We further examine the efficacy of response strategies (polite refusal versus explicit rebuttal) in curbing the occurrence of undesired behavior in subsequent outputs of multi-turn interactions. Our findings reveal that explicit rebuttals significantly outperform polite refusals in preventing the continuation of undesired outputs and nearly eliminate reason-based deception, challenging current practices in model fine-tuning. Accordingly, the two key contributions of this paper are (1) defining and studying reason-based deception, a new type of hidden behavior, and (2) demonstrating that rebuttals provide a more robust response model to harmful requests than refusals, thereby highlighting the need to reconsider the response strategies in fine-tuning approaches.