# Language Models use Lookbacks to Track Beliefs

arXiv: https://arxiv.org/abs/2505.14685 · HTML: https://mlsanity.com/paper/2505.14685/

## Key facts

- Cited by 39 papers in the map (data through 14 Oct 2026).
- Builds on 44 of its 63 references that are in the map.
- First posted on arXiv 20 May 2025.

## Builds on (most cited first, 30 of 44)

- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [In-context Learning and Induction Heads](https://mlsanity.com/paper/2209.11895/) (2022)
- [Probing Classifiers: Promises, Shortcomings, and Advances](https://mlsanity.com/paper/2102.12452/) (2021)
- [Dissecting Recall of Factual Associations in Auto-Regressive Language Models](https://mlsanity.com/paper/2304.14767/) (2023)
- [Causal Abstractions of Neural Networks](https://mlsanity.com/paper/2106.02997/) (2021)
- [Large Language Models Fail on Trivial Alterations to Theory-of-Mind Tasks](https://mlsanity.com/paper/2302.08399/) (2023)
- [Evaluating Large Language Models in Theory of Mind Tasks](https://mlsanity.com/paper/2302.02083/) (2023)
- [Causal Abstraction: A Theoretical Foundation for Mechanistic Interpretability](https://mlsanity.com/paper/2301.04709/) (2023)
- [Implicit Representations of Meaning in Neural Language Models](https://mlsanity.com/paper/2106.00737/) (2021)
- [Understanding Social Reasoning in Language Models with Language Models](https://mlsanity.com/paper/2306.15448/) (2023)
- [FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions](https://mlsanity.com/paper/2310.15421/) (2023)
- [Clever Hans or Neural Theory of Mind? Stress Testing Social Reasoning in Large Language Models](https://mlsanity.com/paper/2305.14763/) (2023)
- [Does Circuit Analysis Interpretability Scale? Evidence from Multiple Choice Capabilities in Chinchilla](https://mlsanity.com/paper/2307.09458/) (2023)
- [Fine-Tuning Enhances Existing Mechanisms: A Case Study on Entity Tracking](https://mlsanity.com/paper/2402.14811/) (2024)
- [Causal Analysis of Syntactic Agreement Mechanisms in Neural Language Models](https://mlsanity.com/paper/2106.06087/) (2021)
- [Inducing Causal Structure for Interpretable Neural Networks](https://mlsanity.com/paper/2112.00826/) (2021)
- [How do Language Models Bind Entities in Context?](https://mlsanity.com/paper/2310.17191/) (2023)
- [Minding Language Models' (Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker](https://mlsanity.com/paper/2306.00924/) (2023)
- [Think Twice: Perspective-Taking Improves Large Language Models' Theory-of-Mind Capabilities](https://mlsanity.com/paper/2311.10227/) (2023)
- [MMToM-QA: Multimodal Theory of Mind Question Answering](https://mlsanity.com/paper/2401.08743/) (2024)
- [How FaR Are Large Language Models From Agents with Theory-of-Mind?](https://mlsanity.com/paper/2310.03051/) (2023)
- [Boosting Theory-of-Mind Performance in Large Language Models via Prompting](https://mlsanity.com/paper/2304.11490/) (2023)
- [How do Decisions Emerge across Layers in Neural Models? Interpretation with Differentiable Masking](https://mlsanity.com/paper/2004.14992/) (2020)
- [HI-TOM: A Benchmark for Evaluating Higher-Order Theory of Mind Reasoning in Large Language Models](https://mlsanity.com/paper/2310.16755/) (2023)
- [LLMs achieve adult human performance on higher-order theory of mind tasks](https://mlsanity.com/paper/2405.18870/) (2024)

## Built on by (most cited first, 30 of 39)

- [Mixing Mechanisms: How Language Models Retrieve Bound Entities In-Context](https://mlsanity.com/paper/2510.06182/) (2025)
- [Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior](https://mlsanity.com/paper/2605.05115/) (2026)
- [Language Model Circuits Are Sparse in the Neuron Basis](https://mlsanity.com/paper/2601.22594/) (2026)
- [Can Aha Moments Be Fake? Towards Quantifying Decorative and True Thinking in Chain-of-Thought](https://mlsanity.com/paper/2510.24941/) (2025)
- [How Causal Abstraction Underpins Computational Explanation](https://mlsanity.com/paper/2508.11214/) (2025)
- [Arithmetic in the Wild: Llama uses Base-10 Addition to Reason About Cyclic Concepts](https://mlsanity.com/paper/2605.01148/) (2026)
- [Mechanistic evaluation of Transformers and state space models](https://mlsanity.com/paper/2505.15105/) (2025)
- [LLMs Process Lists With General Filter Heads](https://mlsanity.com/paper/2510.26784/) (2025)
- [The Story is Not the Science: Execution-Grounded Evaluation of Mechanistic Interpretability Research](https://mlsanity.com/paper/2602.18458/) (2026)
- [Do Language Models Track Entities Across State Changes?](https://mlsanity.com/paper/2605.30233/) (2026)
- [In-Context Algebra](https://mlsanity.com/paper/2512.16902/) (2025)
- [Mechanisms of Prompt-Induced Hallucination in Vision-Language Models](https://mlsanity.com/paper/2601.05201/) (2026)
- [Do explanations generalize across large reasoning models?](https://mlsanity.com/paper/2601.11517/) (2026)
- [Personalization Increases Affective Alignment but Has Role-Dependent Effects on Epistemic Independence in LLMs](https://mlsanity.com/paper/2603.00024/) (2026)
- [Seeing to Generalize: How Visual Data Corrects Binding Shortcuts](https://mlsanity.com/paper/2602.15183/) (2026)
- [The Dual Mechanisms of Spatial Variable Binding in Vision-Language Models](https://mlsanity.com/paper/2603.22278/) (2026)
- [Decomposing Query-Key Feature Interactions Using Contrastive Covariances](https://mlsanity.com/paper/2602.04752/) (2026)
- [The Topological Trouble With Transformers](https://mlsanity.com/paper/2604.17121/) (2026)
- [Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space](https://mlsanity.com/paper/2605.12412/) (2026)
- [A retrieval conditioned rebinding circuit for dynamic entity tracking in large language models](https://mlsanity.com/paper/2606.08644/) (2026)
- [HalluWorld: A Controlled Benchmark for Hallucination via Reference World Models](https://mlsanity.com/paper/2605.19341/) (2026)
- [When Should Models Change Their Minds? Contextual Belief Management in Large Language Models](https://mlsanity.com/paper/2605.30219/) (2026)
- [When Does Activation Steering Change What a Model Computes From?](https://mlsanity.com/paper/2606.29522/) (2026)
- [Compute Globally, Materialize Locally: The Memory Contract of Sparse Event-KV](https://mlsanity.com/paper/2607.23693/) (2026)
- [Artificial Phantasia: Emergent Mental Imagery in Large Language Models](https://mlsanity.com/paper/2509.23108/) (2025)
- [Constructive Circuit Amplification: Improving Math Reasoning in LLMs via Targeted Sub-Network Updates](https://mlsanity.com/paper/2512.16914/) (2025)
- [PDDL-Mind: Large Language Models are Capable on Belief Reasoning with Reliable State Tracking](https://mlsanity.com/paper/2604.17819/) (2026)
- [Who Says What: Symbolic Trimodal Binding Mechanisms in Audio-Visual LLMs](https://mlsanity.com/paper/2609.31193/) (2026)
- [Circuit Distillation](https://mlsanity.com/paper/2509.25002/) (2025)
- [One mechanism for many mental spaces: a shared router over a value slot in language models](https://mlsanity.com/paper/2607.10248/) (2026)

## Abstract (arXiv)

How do language models (LMs) represent characters' beliefs, especially when those beliefs may differ from reality? This question lies at the heart of understanding the Theory of Mind (ToM) capabilities of LMs. We analyze LMs' ability to reason about characters' beliefs using causal mediation and abstraction. We construct a dataset, CausalToM, consisting of simple stories where two characters independently change the state of two objects, potentially unaware of each other's actions. Our investigation uncovers a pervasive algorithmic pattern that we call a lookback mechanism, which enables the LM to recall important information when it becomes necessary. The LM binds each character-object-state triple together by co-locating their reference information, represented as Ordering IDs (OIs), in low-rank subspaces of the state token's residual stream. When asked about a character's beliefs regarding the state of an object, the binding lookback retrieves the correct state OI and then the answer lookback retrieves the corresponding state token. When we introduce text specifying that one character is (not) visible to the other, we find that the LM first generates a visibility ID encoding the relation between the observing and the observed character OIs. In a visibility lookback, this ID is used to retrieve information about the observed character and update the observing character's beliefs. Our work provides insights into belief tracking mechanisms, taking a step toward reverse-engineering ToM reasoning in LMs.