# Gated Activation Steering for Reducing Sycophancy & Hallucination in Medical Question Answering

arXiv: https://arxiv.org/abs/2608.23666 · HTML: https://mlsanity.com/paper/2608.23666/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Hallucination detection from model internals, one of 13 directions in Mechanistic interpretability.
- Builds on 13 of its 18 references that are in the map.
- First posted on arXiv 24 Aug 2026.

## Placed in

- Mechanistic interpretability / [Hallucination detection from model internals](https://mlsanity.com/mechinterp/hallucination-detection-from-model-internals/)

## Builds on (most cited first, 13 of 13)

- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Large Language Models Encode Clinical Knowledge](https://mlsanity.com/paper/2212.13138/) (2022)
- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [Representation Engineering: A Top-Down Approach to AI Transparency](https://mlsanity.com/paper/2310.01405/) (2023)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Domain-Specific Language Model Pretraining for Biomedical Natural Language Processing](https://mlsanity.com/paper/2007.15779/) (2020)
- [Programming Refusal with Conditional Activation Steering](https://mlsanity.com/paper/2409.05907/) (2024)
- [A Question-Entailment Approach to Question Answering](https://mlsanity.com/paper/1901.08079/) (2019)
- [Semantics-Adaptive Activation Intervention for LLMs via Dynamic Steering Vectors](https://mlsanity.com/paper/2410.12299/) (2024)
- [EchoBench: Benchmarking Sycophancy in Medical Large Vision-Language Models](https://mlsanity.com/paper/2509.20146/) (2025)
- [MedInsight: A Multi-Source Context Augmentation Framework for Generating Patient-Centric Medical Responses using Large Language Models](https://mlsanity.com/paper/2403.08607/) (2024)
- [CLINICSUM: Utilizing Language Models for Generating Clinical Summaries from Patient-Doctor Conversations](https://mlsanity.com/paper/2412.04254/) (2024)
- [The Problem of Atypicality in LLM-Powered Psychiatry](https://mlsanity.com/paper/2508.06479/) (2025)

## Abstract (arXiv)

Sycophancy and hallucination are persistent failure modes of Large Language Models (LLMs) across domains. However, it becomes particularly consequential in clinical question answering, where responses must remain grounded in the provided context and robust to user pressure. Hallucination can introduce information that is unsupported by the context, while sycophancy can cause a model to abandon a previously correct answer when challenged by the user. Existing approaches, such as prompt-based safeguards and always-on activation steering, often address these behaviors separately or apply interventions broadly across turns, which can unnecessarily deteriorate responses that were already correct. To address these limitations within a single framework, we employ Inference Time Intervention (ITI) to jointly control both behaviors by learning separate steering directions for hallucination and sycophancy from contrastive clinical pairs and applying them to causally verified attention heads. During runtime, behavior-specific gates then determine when intervention is needed: the hallucination component mitigates unsupported claims, while the sycophancy component mitigates answer shifts caused by user pressure. We evaluate this framework on clinical questions grounded in EHR data while keeping the model weights frozen. Across all evaluation settings, we conducted 15,900 model-response runs. Across 600 pressure trajectories for the 4-billion-parameter model, the unsteered model caved in 570 cases. At the same time, gated steering helped it last longer in 551 of them. It held its ground under pressure at levels comparable to those of models with more than 100 billion parameters, showing that targeted inference-time steering can improve robustness without intervening at every turn.