NeurIPS 2026 In Progress
On the Reasoning Gaps of Large Language Models: A Formal Characterization
Formally characterizing reasoning failures via computational complexity
PAPERS
Publications and preprints from Deepwork Research
Papers in progress. First submissions targeting NeurIPS 2026 and ACL 2027.
Upcoming
ACL 2027 In Progress
A Taxonomy of Failure Modes in LLM-Based Autonomous Agents
Comprehensive survey categorizing agent failure modes