MINT: Evaluating LLMs in Multi-turn Interaction with Tools and Language Feedback
This addresses the discrepancy between research benchmarks and real-world use cases for LLMs, particularly benefiting the open-source community by providing a reproducible evaluation framework for multi-turn interactions.
The authors tackled the problem of evaluating large language models (LLMs) in multi-turn interactions with tools and language feedback, which is often overlooked in single-turn benchmarks, by introducing the MINT benchmark. They found that LLMs generally benefit from tools and feedback with performance gains of 1-8% per tool turn and 2-17% with feedback, but supervised instruction-finetuning and reinforcement learning from human feedback often hurt multi-turn capabilities.
To solve complex tasks, large language models (LLMs) often require multiple rounds of interactions with the user, sometimes assisted by external tools. However, current evaluation protocols often emphasize benchmark performance with single-turn exchanges, neglecting the nuanced interactions among the user, LLMs, and external tools, while also underestimating the importance of natural language feedback from users. These oversights contribute to discrepancies between research benchmark evaluations and real-world use cases. We introduce MINT, a benchmark that evaluates LLMs' ability to solve tasks with multi-turn interactions by (1) using tools and (2) leveraging natural language feedback. To ensure reproducibility, we provide an evaluation framework where LLMs can access tools by executing Python code and receive users' natural language feedback simulated by GPT-4. We repurpose a diverse set of established evaluation datasets focusing on reasoning, coding, and decision-making and carefully curate them into a compact subset for efficient evaluation. Our analysis of 20 open- and closed-source LLMs offers intriguing findings. (a) LLMs generally benefit from tools and language feedback, with performance gains (absolute, same below) of 1-8% for each turn of tool use and 2-17% with natural language feedback. (b) Better single-turn performance does not guarantee better multi-turn performance. (c) Surprisingly, on the LLMs evaluated, supervised instruction-finetuning (SIFT) and reinforcement learning from human feedback (RLHF) generally hurt multi-turn capabilities. We expect MINT can help measure progress and incentivize research in improving LLMs' capabilities in multi-turn interactions, especially for open-source communities where multi-turn human evaluation can be less accessible compared to commercial LLMs with a larger user base.