Chuqin Geng

Chuqin Geng is a computer science researcher at McGill University and Mila working with Xujie Si on neural network verification and mechanistic interpretability. His work examines whether the optimization objectives used to extract sparse computational subgraphs from transformers actually recover the mechanisms the intact model uses.

As first author of the October 2026 preprint “Objective-Level Recovery Gaps in Mechanistic Interpretability: Context Distortion Can Misrank Circuits” (arXiv:2610.02098), Geng isolated two distinct failure modes, context distortion (replacing excluded signals alters the computational inputs on which retained components operate) and the objective-level recovery gap (where the validation faithfulness metric itself rewards a behaviorally inferior circuit of the exact same size). Across 100 persistent KL misrankings from automated discovery pools (EAP, EAP-IG, ACDC, and Edge-SP), Geng showed that partial restoration of recipient-intact context signals repairs the ranking in 96 cases.

Geng’s profile is Builder, covered for turning circuit-evaluation reliability into a controlled, size-matched experimental audit.

Known for. Objective-level recovery gaps and context distortion in mechanistic interpretability circuit evaluation, neural network verification

Coverage on this site

• Objective-Level Recovery Gaps in Mechanistic Interpretability and SAEs October 2026

Related researchers

• XS Xujie SiDepartment of Computer Science, University of Toronto. Mila - Quebec AI InstituteObjective-level recovery gaps and context distortion in mechanistic interpretability circuit discovery, program synthesis and neuro-symbolic verification Builder1 article
• SL Salima LamsiyahDepartment of Computer Science, University of LuxembourgCritical survey of 138 Sparse Autoencoder (SAE) studies for mechanistic interpretability in NLP, four-outcome separation of recoverability, semantic interpretability, mechanistic faithfulness, and behavioral utility Builder1 article
• PB Patrick ButlinEleos AI Research. Previously the Global Priorities Institute, OxfordLead author of the indicator properties framework for assessing consciousness in AI systems AI welfare24 articles
• VA Vikas AgnihotriIndependent Aerospace and Neuromorphic Systems ResearcherNeuromorphic Satellite Intelligence framework, radiation-induced energy overhead model for spiking neural hardware, constrained on-orbit STDP Builder1 article
• MA Matt AngleParadromicsFounder and CEO of Paradromics, developer of the Connexus Direct Data Interface high-data-rate brain-computer interface Builder1 article
• IB Isabella R. BeckettCoauthor of the MaleCNS preprint comparing male and female Drosophila brain connectomes Builder1 article

Chuqin Geng is listed with the other builders, who hold that the claim is about what the hardware does, and the philosophy is left open. The positions themselves are set out in the index of consciousness theories. Every researcher covered on this site is indexed in the directory of consciousness researchers.