Papers by Yichang Xu
H3Fusion: Helpful, Harmless, Honest Fusion of Aligned LLMs (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing approaches to align pre-trained LLMs with instructions for one property are difficult to fine-tune. |
| Approach: | They propose a mixture-of-experts-based fusion mechanism that models alignment as a controllable drift within the subspace, guided by a drift-regularization loss to balance competing alignment dimensions. |
| Outcome: | Extensive evaluations of three benchmark datasets show that H3Fusion outperforms each individually aligned model by 11.37% and provides stronger robustness compared to the state-of-the-art LLM ensemble approaches by 13.77% and model-merging approaches by 6.18 %. |
Confidence v.s. Critique: A Decomposition of Self-Correction Capability for LLMs (2025.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to improve self-correction performance of Large Language Models are based on intrinsic selfcorrectione, which allows the model to check and revise its selfgenerated answers without external feedback. |
| Approach: | They propose to decompose the self-correction capability into confidence and critique capabilities and a metric for overall self-corretion capability evaluation. |
| Outcome: | The proposed method outperforms vanilla SFT and achieves much higher accuracy after self-correction. |
A Probabilistic Inference Scaling Theory for LLM Self-Correction (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have demonstrated the capability to refine their generated answers through self-correction, enabling continuous performance improvement over multiple rounds. |
| Approach: | They propose a probabilistic theory to model the dynamics of accuracy change and explain performance improvements observed in multi-round self-correction. |
| Outcome: | The proposed model can predict accuracy curves and improve accuracy over multiple rounds. |