Shanda Li

Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces

FrontierCO: Real-World and Large-Scale Evaluation of Machine Learning Solvers for Combinatorial Optimization

CoMind: Towards Community-Driven Agents for Machine Learning Engineering

Sample Complexity and Representation Ability of Test-time Scaling Paradigms

CodePDE: An Inference Framework for LLM-driven PDE Solver Generation

TFG-Flow: Training-free Guidance in Multimodal Generative Flow

Maximal Update Parametrization and Zero-Shot Hyperparameter Transfer for Fourier Neural Operators

Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for LLM Problem-Solving