arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2501.09891cs.AI
收藏

演进更深的大语言模型思考

Evolving Deeper LLM Thinking

Google DeepMind(谷歌DeepMind);UC San Diego(加州大学圣迭戈分校);University of Alberta(阿尔伯塔大学)

Kuang-Huei Lee, Ian Fischer, Yueh-Hua Wu, Dave Marwood, Shumeet Baluja, Dale Schuurmans, Xinyun Chen

AI总结:

本研究提出Mind Evolution进化搜索策略,通过语言模型生成、重组、优化候选响应以扩展大模型推理时算力,在控制成本下于自然语言规划任务显著优于同类策略,在两项基准上用Gemini 1.5 Pro达成超98%的求解率。

摘要:

我们探索了一种用于扩展大语言模型推理时计算量的进化搜索策略。所提出的Mind Evolution(心智演进)方法利用语言模型生成、重组并优化候选响应。该方法无需在有解决方案评估器时将底层推理问题形式化。在控制推理成本的条件下,我们发现在自然语言规划任务中,Mind Evolution的表现显著优于Best-of-N(最优N采样)、Sequential Revision(序列修订)等其他推理策略。在TravelPlanner和Natural Plan基准测试中,使用Gemini 1.5 Pro的Mind Evolution无需借助形式化求解器,即可解决超过98%的问题实例。

英文摘要:

We explore an evolutionary search strategy for scaling inference time compute in Large Language Models. The proposed approach, Mind Evolution, uses a language model to generate, recombine and refine candidate responses. The proposed approach avoids the need to formalize the underlying inference problem whenever a solution evaluator is available. Controlling for inference cost, we find that Mind Evolution significantly outperforms other inference strategies such as Best-of-N and Sequential Revision in natural language planning tasks. In the TravelPlanner and Natural Plan benchmarks, Mind Evolution solves more than 98% of the problem instances using Gemini 1.5 Pro without the use of a formal solver.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑