🦅 State-of-the-Art Prompt Engineering
🎵 You say the coast is clear, but you won't catch me out 🎵
AI @ UIUC is proud to announce that a research project led by our Principal Investigator Andy Zhou has officially broken state-of-the-art in LLM problem-solving benchmarks!
Paper: Language Agent Tree Search Unifies Reasoning Acting and Planning in Language Models
Project Page: https://andyz245.github.io/LanguageAgentTreeSearch/
🤖 Additionally, our technical staff has developed an open-source tech demo to display the method’s capability in code_generation: CodeLATS
With GPT-4, LATS hits state-of-the-art on HumanEval for programming with a score of 94.4%!
Congratulations to Andy Zhou and the team for this amazing accomplishment! More publications to come very soon :)
Local Artist Spotlight of the Day (lindsay.artssss) @lindsay.artssss
