absorb.md

Instructional Segment Embeddings Ise

1 mentions across 0 people

Unknown speaker
paper · 2026-06-10
Recommended

“Combined with hierarchy-specific delimiter tokens (Chen et al., 2025) and Instructional Segment Embeddings (ISE; Wu et al., 2025), GW-DPO with the bilateral schedule Pareto-improves over standard DPO and the linear variant on Llama-3.1-8B-Instruct, raising macro pairwise priority adherence while keeping over-refusal at half the standard DPO rate.”

Gravity-Weighted DPO: Enforcing Multi-Level Instruction Hierarchies in LLMs ↗