hello, i'm wesley. here's some of my publicized work
- Benchmark: Can LLMs Optimize Code Compilers Can't?
- RECKONING: A Counterfactual Benchmark for Post-Break Metacognitive Recovery in Long-Horizon Agents
- Dense vs. MoE Layer Specialization during RL in Interleaved Models
- Building an LM from Scratch
- Every Sutton & Barto RL Algorithm, Applied to Gomoku
- LLM Utility Learning from Perceived Outcomes (OM-LEU)
- ML-Driven Market-Neutral Pairs Trading


