Can I read 2025.findings Acl.958 on EtoBox?
2025.findings Acl.958 by mingashuatod is a document available to read on EtoBox.
What is 2025.findings Acl.958 about?
MultiChallenge is a new benchmark designed to evaluate large language models (LLMs) on their ability to conduct realistic multi-turn conversations with human users, focusing on four key challenge categories: instruction retention, inference memory, reliable versioned editing, and self-coherence. Despite achieving high scores on existing benchmarks, frontier LLMs struggle with MultiChallenge, with the best model scoring only 41.4% accuracy. The benchmark employs a hybrid evaluation method combining human and
- Author
- mingashuatod
- Language
- EN