Jiashu He, Meizhu Liu, Olaitan P Olaleye, Amit Agarwal, M. Avendi, Yassi Abbasi, Matthew Rowe, Hitesh Laxmichand Patel, Paul Li, Tao Sheng, Sujith Ravi, Dan Roth
View original ↗Implement an entropy-guided decoding strategy that optimizes LLM reasoning by monitoring token probabilities. This provides a low-overhead alternative to computationally expensive self-consistency chains.
Suggested repo: entrope-decode
"Boost LLM reasoning efficiency with entropy-guided decoding."
Estimated effort: 50h