# TD-DPO: Difference-Aware Preference Optimization for Mitigating Sycophancy in Clinical Autism Intervention Dialogue

arXiv: https://arxiv.org/abs/2607.18304 · HTML: https://mlsanity.com/paper/2607.18304/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Medical LLM alignment, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 30 of its 55 references that are in the map.
- First posted on arXiv 17 Jul 2026.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Medical LLM alignment](https://mlsanity.com/alignment/medical-llm-alignment/)

## Builds on (most cited first, 30 of 30)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation](https://mlsanity.com/paper/1609.08144/) (2016)
- [Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models](https://mlsanity.com/paper/2506.05176/) (2025)
- [A Diversity-Promoting Objective Function for Neural Conversation Models](https://mlsanity.com/paper/1510.03055/) (2015)
- [SimPO: Simple Preference Optimization with a Reference-Free Reward](https://mlsanity.com/paper/2405.14734/) (2024)
- [ORPO: Monolithic Preference Optimization without Reference Model](https://mlsanity.com/paper/2403.07691/) (2024)
- [Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation](https://mlsanity.com/paper/2401.08417/) (2024)
- [SWIFT:A Scalable lightWeight Infrastructure for Fine-Tuning](https://mlsanity.com/paper/2408.05517/) (2024)
- [Efficient and Effective Text Encoding for Chinese LLaMA and Alpaca](https://mlsanity.com/paper/2304.08177/) (2023)
- [Self-Play Preference Optimization for Language Model Alignment](https://mlsanity.com/paper/2405.00675/) (2024)
- [Zhongjing: Enhancing the Chinese Medical Capabilities of Large Language Model through Expert Feedback and Real-world Multi-turn Dialogue](https://mlsanity.com/paper/2308.03549/) (2023)
- [Measuring Sycophancy of Language Models in Multi-turn Dialogues](https://mlsanity.com/paper/2505.23840/) (2025)
- [CPsyCoun: A Report-based Multi-turn Dialogue Reconstruction and Evaluation Framework for Chinese Psychological Counseling](https://mlsanity.com/paper/2405.16433/) (2024)
- [When Large Language Models contradict humans? Large Language Models' Sycophantic Behaviour](https://mlsanity.com/paper/2311.09410/) (2023)
- [Anchored Preference Optimization and Contrastive Revisions: Addressing Underspecification in Alignment](https://mlsanity.com/paper/2408.06266/) (2024)
- [CBT-LLM: A Chinese Large Language Model for Cognitive Behavioral Therapy-based Mental Health Question Answering](https://mlsanity.com/paper/2403.16008/) (2024)
- [TIS-DPO: Token-level Importance Sampling for Direct Preference Optimization With Estimated Weights](https://mlsanity.com/paper/2410.04350/) (2024)
- [Be Friendly, Not Friends: How LLM Sycophancy Shapes User Trust](https://mlsanity.com/paper/2502.10844/) (2025)
- [TRUTH DECAY: Quantifying Multi-Turn Sycophancy in Language Models](https://mlsanity.com/paper/2503.11656/) (2025)
- [Selective Preference Optimization via Token-Level Reward Function Estimation](https://mlsanity.com/paper/2408.13518/) (2024)
- [Sycophancy under Pressure: Evaluating and Mitigating Sycophantic Bias via Adversarial Dialogues in Scientific QA](https://mlsanity.com/paper/2508.13743/) (2025)
- [T-REG: Preference Optimization with Token-Level Reward Regularization](https://mlsanity.com/paper/2412.02685/) (2024)
- [ConfPO: Exploiting Policy Model Confidence for Critical Token Selection in Preference Optimization](https://mlsanity.com/paper/2506.08712/) (2025)
- [Optimal Transport-Based Token Weighting scheme for Enhanced Preference Optimization](https://mlsanity.com/paper/2505.18720/) (2025)
- [ASD-Chat: An Innovative Dialogue Intervention System for Children with Autism based on LLM and VB-MAPP](https://mlsanity.com/paper/2409.01867/) (2024)
- [Echo-Teddy: Preliminary Design and Development of Large Language Model-based Social Robot for Autistic Students](https://mlsanity.com/paper/2502.04029/) (2025)
- [From Synthesis to Clinical Assistance: A Strategy-Aware Agent Framework for Autism Intervention based on Real Clinical Dataset](https://mlsanity.com/paper/2605.02916/) (2026)

## Abstract (arXiv)

The sycophancy of large language models can increase the safety risk in intervention dialogue for autistic children. Supervised fine-tuning can somewhat reduce sycophancy, but relying solely on positive examples is often insufficient to identify and correct failure patterns. We observe that sycophancy behaviors can often be localized to a limited span within the model response. In this regime, sequence-level preference optimization can over-update preference-irrelevant tokens and degrade intervention ability. To address this, we propose the Minimal Edit Data Augmentation (MEDA) strategy to construct controlled, stable, minimal edit preference pairs and Token-level Difference Direct Preference Optimization (TD-DPO), which upweights difference tokens between chosen and rejected responses while downweighting shared tokens to suppress background drift. Extensive experiments across multiple backbones and evaluators show that TD-DPO achieves a better trade-off between sycophancy mitigation and intervention ability retention in our offline settings, highlighting its potential as a practical alignment approach for autism intervention.