Showing 2026Show all
2 papers · 1 filter
cs.CL2026
Instruction Quality Matters: Refining Instructions for Effective Preference Learning
Seohyeong Lee, Hwaran Lee, Buru Chang
Preference learning optimizes models using response pairs, yet the informativeness of these pairs is fundamentally shaped by the instructions from which they are generated. We iden…
cs.CL2026
HarDBench: A Benchmark for Draft-Based Co-Authoring Jailbreak Attacks for Safe Human-LLM Collaborative Writing
Euntae Kim, Soomin Han, Buru Chang
Large language models (LLMs) are increasingly used as co-authors in collaborative writing, where users begin with rough drafts and rely on LLMs to complete, revise, and refine thei…