# Causal Stories from Sensor Traces: Auditing Epistemic Overreach in LLM-Generated Personal Sensing Explanations

arXiv: https://arxiv.org/abs/2605.08590 · HTML: https://mlsanity.com/paper/2605.08590/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 78 references that are in the map.
- First posted on arXiv 9 May 2026.

## Builds on (most cited first, 21 of 21)

- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment](https://mlsanity.com/paper/2303.16634/) (2023)
- [Language Models (Mostly) Know What They Know](https://mlsanity.com/paper/2207.05221/) (2022)
- [A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions](https://mlsanity.com/paper/2311.05232/) (2023)
- [FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation](https://mlsanity.com/paper/2305.14251/) (2023)
- [Explanation in Artificial Intelligence: Insights from the Social Sciences](https://mlsanity.com/paper/1706.07269/) (2017)
- [Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting](https://mlsanity.com/paper/2305.04388/) (2023)
- [On Faithfulness and Factuality in Abstractive Summarization](https://mlsanity.com/paper/2005.00661/) (2020)
- [DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models](https://mlsanity.com/paper/2306.11698/) (2023)
- [QAFactEval: Improved QA-Based Factual Consistency Evaluation for Summarization](https://mlsanity.com/paper/2112.08542/) (2021)
- [A Dataset for Answering Time-Sensitive Questions](https://mlsanity.com/paper/2108.06314/) (2021)
- [TimeBench: A Comprehensive Evaluation of Temporal Reasoning Abilities in Large Language Models](https://mlsanity.com/paper/2311.17667/) (2023)
- [Personalized Machine Learning for Robot Perception of Affect and Engagement in Autism Therapy](https://mlsanity.com/paper/1802.01186/) (2018)
- [MindScape Study: Integrating LLM and Behavioral Sensing for Personalized AI-Driven Journaling Experiences](https://mlsanity.com/paper/2409.09570/) (2024)
- [PhysioLLM: Supporting Personalized Health Insights with Wearables and Large Language Models](https://mlsanity.com/paper/2406.19283/) (2024)
- [SensorChat: Answering Qualitative and Quantitative Questions during Long-Term Multimodal Sensor Interactions](https://mlsanity.com/paper/2502.02883/) (2025)
- [RubRIX: Rubric-Driven Risk Mitigation in Caregiver-AI Interactions](https://mlsanity.com/paper/2601.13235/) (2026)
- [MindGuard: Towards Accessible and Sitgma-free Mental Health First Aid via Edge LLM](https://mlsanity.com/paper/2409.10064/) (2024)
- [PAIR-SAFE: A Paired-Agent Approach for Runtime Auditing and Refining AI-Mediated Mental Health Support](https://mlsanity.com/paper/2601.12754/) (2026)
- [Imputation Matters: A Deeper Look into an Overlooked Step in Longitudinal Health and Behavior Sensing Research](https://mlsanity.com/paper/2412.06018/) (2024)

## Built on by (most cited first, 2 of 2)

- [Not All Nudges Land: Behavioral Controllability and Elaboration Quality in AI-Supported Journaling](https://mlsanity.com/paper/2608.12582/) (2026)
- [Personalizing Personal Health Interfaces: Co-Design with Generative AI](https://mlsanity.com/paper/2609.15046/) (2026)

## Abstract (arXiv)

LLMs are increasingly used to explain personal sensing data, translating traces of activity and mood into natural-language accounts of why an anomalous day may have occurred. However, such explanations can sound coherent and personally meaningful even when the underlying evidence is sparse or missing. We introduce epistemic overreach (EO) as a measure for cases where a generated explanation implies more than the available sensing evidence can justify. To audit how often and in what forms EO occurs, we obtained anomalous-day scenarios from three longitudinal sensing datasets of college students: StudentLife, GLOBEM, and CollegeExperience. Across activity, sleep, and affect anomalies, we generated 14,922 explanations using three LLM families -- Llama, Qwen, and GPT -- under two prompting conditions: one minimally constrained prompt and another prompt explicitly instructing models to bound claims to the data. For each scenario, we varied the amount of behavioral evidence available to the model to examine whether more evidence reduces EO. We evaluated each explanation using a structured rubric, decomposing EO into the dimensions of unsupported causal attribution, unacknowledged data gaps, overconfident language, temporal inconsistency, and diagnostic inference. We find that LLMs routinely attribute anomalous days to causes without sufficient support from the data, and that this pattern replicates across datasets, anomaly types, and model families. Further, providing richer context does not reliably reduce EO; bounded prompting helps but does not eliminate it. These findings suggest that evidential grounding should be a first-order evaluation criterion for LLM-generated personal sensing explanations, alongside fluency and plausibility. We argue that personal sensing explanations require evidential discipline: systems must distinguish what is observed, what is inferred, and what remains unknown.