List your company

Multi-turn Interaction-based Problem Solving on MINT-Bench 1.0 (test)

11.76Code Generation ScoreLLAMA PRO - INSTRUCT
-0.47042.70485.889.0552Jan 4, 2024

Evaluation results

MethodLinks
LLAMA PRO - INSTRUCTinteraction turns (k)=52024.0111.7629.19.8114.68
Mistral-Instruct-v0.1interaction turns (k)=52024.016.6234.338.5413.99
CodeLLaMA-7B-Instructinteraction turns (k)=52024.012.2117.167.918.7
AgentLM-7Binteraction turns (k)=52024.011.479.78.867.34
LLaMA2-7B-Chatinteraction turns (k)=52024.010013.617.34