Jun 12, 02:12 AM

$Google and Researchers Enhance Mathematical Reasoning in LLMs, Achieving 69.4% on MATH Benchmarks$

Google and Researchers Enhance Mathematical Reasoning in LLMs, Achieving 69.4% on MATH Benchmarks

Recent advancements in large language models (LLMs) have shown mixed results in mathematical reasoning tasks. A new benchmark reveals that while LLMs excel in straightforward math problems, they struggle with creative and multi-step questions, even with the use of chain-of-thought (CoT) prompting. Google's latest project, 'Improve Mathematical Reasoning in Language Models by Automated Process Supervision,' employs Monte Carlo Tree Search (MCTS) to enhance the collection of high-quality process supervision data, resulting in a significant performance boost from 51% to 69.4% on MATH benchmarks without human intervention. Additionally, researchers from Fudan University and The Hong Kong Polytechnic University introduced a method to access GPT-4 level Mathematical Olympiad solutions using Monte Carlo Tree Self-refine with LLaMa-3 8B, significantly improving success rates across MATH and Olympiad-level benchmarks.

#Google #Monte Carlo Tree Search #Fudan University #GPT #Mathematical Olympiad #LLaMa #Olympiad

Written with ChatGPT (GPT-4o).

Sources

Additional media

$Image #1 for story google-researchers-enhance-mathematical-reasoning-llms-achieving-69-4-on-math$

$Image #2 for story google-researchers-enhance-mathematical-reasoning-llms-achieving-69-4-on-math$

$Image #3 for story google-researchers-enhance-mathematical-reasoning-llms-achieving-69-4-on-math$