Fangzhi Xu

ENVISIONS. [ACL 2025] A Neural-Symbolic Self-Training Framework

phi-Decoding. [ACL 2025] An inference-time decoding strategy with adaptive foresight sampling

Genius. [ACL 2025] A Generalizable and Purely Unsupervised Self-Training Framework

Symbol-LLM. [ACL 2024] The project of Symbol-LLM

Odyssey-Arena. Extremely Long-Horizon Agentic Tasks Requiring Active Acting and Inductive Reasoning

Logiformer. [SIGIR 2022] The implementation of Logiformer

TaCo. [IEEE TNNLS] The benchmark ZsLR and the implementation of TaCo

NLP_HW2. The second Homework of NLP

MoCA. [Pattern Recognition] The implementation of MoCA

NLP_Final. NLP Final Project：SQuAD

PathReasoner. [ACL 2024] The implementation of PathReasoner

NCISurvey. Neural Code Intelligence Survey 2024

ReasoningPaperList. Paper List for recent works on reasoning

SeeClick. The model, data and code for the visual GUI Agent SeeClick