Not All LLM Reasoners Are Created Equal

[Free GPT-4]
[DeepSeek]
[PDF] arxiv.org

Multi-Step Reasoning in Korean and the Emergent Mirage

G Son, H Ko, D Choi - arxiv preprint arxiv:2501.05712, 2025 - arxiv.org
We introduce HRMCR (HAE-RAE Multi-Step Commonsense Reasoning), a benchmark
designed to evaluate large language models' ability to perform multi-step reasoning in …
Enregistrer Citer Autres articles Les 2 versionsFree GPT-4 DeepSeek Version HTML