# CoSToM:Causal-oriented Steering for Intrinsic Theory-of-Mind Alignment in Large Language Models

Source: [CoSToM:Causal-oriented Steering for Intrinsic Theory-of-Mind Alignment in Large Language Models](<https://www.sota2.com/research/paper/69ddc9ed7d0172140ec79860>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-15T05:00:04.591Z

Authors: Mengfan Li, Xuanhua Shi, Yang Deng

Published: 2026-04-11T05:04:37.000Z

## Abstract

Theory of Mind (ToM), the ability to attribute mental states to others, is a hallmark of social intelligence. While large language models (LLMs) demonstrate promising performance on standard ToM benchmarks, we observe that they often fail to generalize to complex task-specific scenarios, relying heavily on prompt scaffolding to mimic reasoning. The critical misalignment between the internal knowledge and external behavior raises a fundamental question: Do LLMs truly possess intrinsic cognition, and can they externalize this internal knowledge into stable, high-quality behaviors? To answer this, we introduce CoSToM (Causal-oriented Steering for ToM alignment), a framework that transitions from mechanistic interpretation to active intervention. First, we employ causal tracing to map the internal distribution of ToM features, empirically uncovering the internal layers' characteristics in encoding fundamental ToM semantics. Building on this insight, we implement a lightweight alignment framework via targeted activation steering within these ToM-critical layers. Experiments demonstrate that CoSToM significantly enhances human-like social reasoning capabilities and downstream dialogue quality.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Dialogue Generation on NEGOTIATIONTOM N=100](<https://www.sota2.com/research/sota/dialogue-generation-on-negotiationtom-n-100.md>): Dialogue Generation; NEGOTIATIONTOM N=100; ToM (Llama-3.3-70B): 0.751.

- [Dialogue Generation on PERSUASIVETOM](<https://www.sota2.com/research/sota/dialogue-generation-on-persuasivetom.md>): Dialogue Generation; PERSUASIVETOM; ToM Score (Judge: Llama-3.3-70B): 80.2.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2604.10031>)

[PDF](<https://arxiv.org/pdf/2604.10031.pdf>)
