Suxin Ji
Titles as the sources gave them. In 1 entry, with 0 checked quotations.
In the Base
Every entry they appear in.
TLRS-DPO: Token-Level Risk Scoring for Better Direct Preference Optimization
According to the abstract, Suxin Ji and Hungtao Wan propose TLRS-DPO, which adds token-level risk scores to DPO training so only high-risk tokens in rejected responses are penalised, and report better alignment and safety with minimal cost to helpfulness.