Claude Instant 1

anchor

Findings

• IC-1330All 20 evaluated LLMs improve in multi-turn task-solving with additional tool-use turns and GPT-4-simulated language feedback
• IC-1361GPT-4 and other state-of-the-art LLMs achieve near-human accuracy in inferring personal attributes from unstructured text
• IC-751Arena-Hard-200 reveals larger performance gaps between open and proprietary LLMs than MT-Bench