Papers by Xiaonan Xu
Long Chain-of-Thought Fine-tuning via Understanding-to-Reasoning Transition (2025.emnlp-main)
Copied to clipboard
Chenxin An, Zhihui Xie, Xiaonan Li, Ming Zhong, Shansan Gong, Lei Li, Jun Zhang, Jingjing Xu, Lingpeng Kong
| Challenge: | Existing research on long-context scaling in language models has focused on managing lengthy input prompts instead of producing long outputs. |
| Approach: | They propose a sequence-level curriculum learning framework that shifts a model’s focus from interpreting long chain-of-thoughts to generating them. |
| Outcome: | Experiments on rigorous reasoning benchmarks, including AIME24 and GPQA Diamond, show that the proposed approach surpasses standard fine-tuning by over 10% while maintaining robust performance on understanding tasks. |
Who did what to Whom? Language models and humans respond diversely to features affecting argument hierarchy construction (2022.aacl-main)
Copied to clipboard
| Challenge: | Pre-trained transformer-based language models have achieved state-of-the-art performance in many areas of NLP. |
| Approach: | They propose to construct argument hierarchy similar to humans' constructs of telicity, agency, and individuation . authors propose to use a Chinese structure to analyze human and transformer-based models' preference for telic context and atelic feature . |
| Outcome: | The proposed model and humans respond to non-)agentive features in telic context and atelic feature very similarly. |