Showing cs.CLShow all
2 papers · 1 filter
cs.CL2024
Distance between Relevant Information Pieces Causes Bias in Long-Context LLMs
Runchu Tian, Yanghao Li, Yuepeng Fu +10
Positional bias in large language models (LLMs) hinders their ability to effectively process long inputs. A prominent example is the "lost in the middle" phenomenon, where LLMs str…
cs.CL2024
TIS-DPO: Token-level Importance Sampling for Direct Preference Optimization With Estimated Weights
Aiwei Liu, Haoping Bai, Zhiyun Lu +9
Direct Preference Optimization (DPO) has been widely adopted for preference alignment of Large Language Models (LLMs) due to its simplicity and effectiveness. However, DPO is deriv…