EPISODE · Aug 23, 2025 · 20 MIN
Deep Think with Confidence
from Daily Paper Cast · host Jingwen Liang, Gengyu Wang
🤗 Upvotes: 26 | cs.LG Authors: Yichao Fu, Xuewei Wang, Yuandong Tian, Jiawei Zhao Title: Deep Think with Confidence Arxiv: http://arxiv.org/abs/2508.15260v1 Abstract: Large Language Models (LLMs) have shown great potential in reasoning tasks through test-time scaling methods like self-consistency with majority voting. However, this approach often leads to diminishing returns in accuracy and high computational overhead. To address these challenges, we introduce Deep Think with Confidence (DeepConf), a simple yet powerful method that enhances both reasoning efficiency and performance at test time. DeepConf leverages model-internal confidence signals to dynamically filter out low-quality reasoning traces during or after generation. It requires no additional model training or hyperparameter tuning and can be seamlessly integrated into existing serving frameworks. We evaluate DeepConf across a variety of reasoning tasks and the latest open-source models, including Qwen 3 and GPT-OSS series. Notably, on challenging benchmarks such as AIME 2025, DeepConf@512 achieves up to 99.9% accuracy and reduces generated tokens by up to 84.7% compared to full parallel thinking.
What this episode covers
🤗 Upvotes: 26 | cs.LG Authors: Yichao Fu, Xuewei Wang, Yuandong Tian, Jiawei Zhao Title: Deep Think with Confidence Arxiv: http://arxiv.org/abs/2508.15260v1 Abstract: Large Language Models (LLMs) have shown great potential in reasoning tasks through test-time scaling methods like self-consistency with majority voting. However, this approach often leads to diminishing returns in accuracy and high computational overhead. To address these challenges, we introduce Deep Think with Confidence (DeepConf), a simple yet powerful method that enhances both reasoning efficiency and performance at test time. DeepConf leverages model-internal confidence signals to dynamically filter out low-quality reasoning traces during or after generation. It requires no additional model training or hyperparameter tuning and can be seamlessly integrated into existing serving frameworks. We evaluate DeepConf across a variety of reasoning tasks and the latest open-source models, including Qwen 3 and GPT-OSS series. Notably, on challenging benchmarks such as AIME 2025, DeepConf@512 achieves up to 99.9% accuracy and reduces generated tokens by up to 84.7% compared to full parallel thinking.
NOW PLAYING
Deep Think with Confidence
No transcript for this episode yet
Similar Episodes
No similar episodes found.