MT-Bench

The 'MT-Bench' is a benchmark designed to evaluate the performance of large language models in controlled settings, focusing on their capabilities and limitations in single-session tasks.

Papers using MT-Bench (6)

ASA: Backbone-Training-Free Representation Engineering for Tool-Calling Agents2026MTR-Bench: A Comprehensive Benchmark for Multi-Turn Reasoning Evaluation2025Evaluating Agentic AI In The Wild: Failure Modes, Drift Patterns, And A Production Evaluation Framework2026MMoA: An AI-Agent framework with recurrence for Memoried Mixure-of-Agent2026CoEvol: Constructing Better Responses for Instruction Finetuning through Multi-Agent Cooperation2024 · 1 citesA System for Morphology-Task Generalization via Unified Representation and Behavior Distillation2022