Decomposing Theory of Mind: How Emotional Processing Mediates ToM Abilities in LLMs

• URL: http://arxiv.org/abs/2511.15895v1
• Date: Wed, 19 Nov 2025 21:56:00 GMT
• Title: Decomposing Theory of Mind: How Emotional Processing Mediates ToM Abilities in LLMs
• Authors: Ivan Chulo, Ananya Joshi,
• Abstract summary: We propose decomposing ToM in language models by comparing steered versus baseline LLMs' activations.<n>We find improved performance on belief attribution tasks (32.5% to 46.7% accuracy) is mediated by activations processing emotional content.<n>This suggests that successful ToM abilities in LLMs are mediated by emotional understanding, not analytical reasoning.
• Score: 6.14481021961242
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recent work shows activation steering substantially improves language models' Theory of Mind (ToM) (Bortoletto et al. 2024), yet the mechanisms of what changes occur internally that leads to different outputs remains unclear. We propose decomposing ToM in LLMs by comparing steered versus baseline LLMs' activations using linear probes trained on 45 cognitive actions. We applied Contrastive Activation Addition (CAA) steering to Gemma-3-4B and evaluated it on 1,000 BigToM forward belief scenarios (Gandhi et al. 2023), we find improved performance on belief attribution tasks (32.5\% to 46.7\% accuracy) is mediated by activations processing emotional content : emotion perception (+2.23), emotion valuing (+2.20), while suppressing analytical processes: questioning (-0.78), convergent thinking (-1.59). This suggests that successful ToM abilities in LLMs are mediated by emotional understanding, not analytical reasoning.

Related papers

• Unveiling the Cognitive Compass: Theory-of-Mind-Guided Multimodal Emotion Reasoning [31.790359663851305]
  genuine affective intelligence requires explicit modeling of Theory of Mind (ToM), the cognitive substrate from which emotions arise.<n>We introduce HitEmotion, a ToM-grounded hierarchical benchmark that diagnoses capability breakpoints across increasing levels of cognitive depth.<n>Second, we propose a ToM-guided reasoning chain that tracks mental states and calibrates cross-modal evidence to achieve faithful emotional reasoning.
  arXiv  Detail & Related papers  (2026-02-01T02:26:12Z)
• Emotion-Coherent Reasoning for Multimodal LLMs via Emotional Rationale Verifier [53.55996102181836]
  We propose the Emotional Rationale Verifier (ERV) and an Explanation Reward.<n>Our method guides the model to produce reasoning that is explicitly consistent with the target emotion.<n>We show that our approach not only enhances alignment between explanation and prediction but also empowers MLLMs to deliver emotionally coherent, trustworthy interactions.
  arXiv  Detail & Related papers  (2025-10-27T16:40:17Z)
• Evaluating & Reducing Deceptive Dialogue From Language Models with Multi-turn RL [64.3268313484078]
  Large Language Models (LLMs) interact with millions of people worldwide in applications such as customer support, education and healthcare.<n>Their ability to produce deceptive outputs, whether intentionally or inadvertently, poses significant safety concerns.<n>We investigate the extent to which LLMs engage in deception within dialogue, and propose the belief misalignment metric to quantify deception.
  arXiv  Detail & Related papers  (2025-10-16T05:29:36Z)
• Customizing Visual Emotion Evaluation for MLLMs: An Open-vocabulary, Multifaceted, and Scalable Approach [29.502292089901825]
  We argue that this inconsistency stems partly from constraints in existing evaluation methods.<n>We propose an Emotion Statement Judgment task that overcomes these constraints.<n>We devise an automated pipeline that efficiently constructs emotion-centric statements with minimal human effort.
  arXiv  Detail & Related papers  (2025-09-26T06:30:39Z)
• MME-Emotion: A Holistic Evaluation Benchmark for Emotional Intelligence in Multimodal Large Language Models [108.61337743051483]
  We present MME-Emotion, a systematic benchmark that assesses both emotional understanding and reasoning capabilities of MLLMs.<n>MME-Emotion contains over 6,000 curated video clips with task-specific questioning-answering (QA) pairs, spanning broad scenarios to formulate eight emotional tasks.<n>It incorporates a holistic evaluation suite with hybrid metrics for emotion recognition and reasoning, analyzed through a multi-agent system framework.
  arXiv  Detail & Related papers  (2025-08-11T03:14:55Z)
• Emotion-LLaMA: Multimodal Emotion Recognition and Reasoning with Instruction Tuning [55.127202990679976]
  We introduce the MERR dataset, containing 28,618 coarse-grained and 4,487 fine-grained annotated samples across diverse emotional categories. This dataset enables models to learn from varied scenarios and generalize to real-world applications. We propose Emotion-LLaMA, a model that seamlessly integrates audio, visual, and textual inputs through emotion-specific encoders.
  arXiv  Detail & Related papers  (2024-06-17T03:01:22Z)
• NegativePrompt: Leveraging Psychology for Large Language Models Enhancement via Negative Emotional Stimuli [21.846500669385193]
  Large Language Models (LLMs) have become integral to a wide spectrum of applications. LLMs possess emotional intelligence, which can be further developed through positive emotional stimuli. We introduce NegativePrompt, a novel approach underpinned by psychological principles.
  arXiv  Detail & Related papers  (2024-05-05T05:06:07Z)
• Emotionally Numb or Empathetic? Evaluating How LLMs Feel Using EmotionBench [83.41621219298489]
  We evaluate Large Language Models' (LLMs) anthropomorphic capabilities using the emotion appraisal theory from psychology. We collect a dataset containing over 400 situations that have proven effective in eliciting the eight emotions central to our study. We conduct a human evaluation involving more than 1,200 subjects worldwide.
  arXiv  Detail & Related papers  (2023-08-07T15:18:30Z)
• Large Language Models Understand and Can be Enhanced by Emotional Stimuli [53.53886609012119]
  We take the first step towards exploring the ability of Large Language Models to understand emotional stimuli. Our experiments show that LLMs have a grasp of emotional intelligence, and their performance can be improved with emotional prompts. Our human study results demonstrate that EmotionPrompt significantly boosts the performance of generative tasks.
  arXiv  Detail & Related papers  (2023-07-14T00:57:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.