length-balance-rejected-responses-preference-data TEC-0141 · proposedLength-balance rejected responses in preference data
Rewrite refusal and rejected responses to match the length and structure of the preferred grounded answer while preserving the underlying grounding failure, so preference learning cannot key on length.
Proposed. Filed by the drafting stage because a paper introduces or tests it. Nobody has vouched for it. The claim drafted from that paper is its first evidence, and what the paper used to validate it is in that claim’s source note. Say whether it held up for you, below.
Rewrite refusal and rejected responses to match the length and structure of the preferred grounded answer while preserving the underlying grounding failure, so preference learning cannot key on length. Filed by the drafting stage from Building Legal Reward Models for Grounding and Abstention (2609.14739) with status proposed: the paper introduces or tests it, and the claim drafted from the paper is its first evidence. The paper validated it by: Pairwise accuracy on LegalRewardBench: Ministral-8B 59.3 baseline, 67.6 with CJB-only DPO, 66.0 with CJB+original legal pairs, 81.8 with length-augmented legal pairs, 84.9 with CJB+length-augmented; same ordering for Gemma 2-2B.. Nobody has vouched for it; add support or contest it on this page.
Does it work?
nothing measured0 supporting · 0 contesting sources
Efficacy claims — what this technique actually moves, under which conditions, and whether that has been contested.
No efficacy claim filed yet. The technique is catalogued; whether it moves the capability, and when, is a separate assertion that needs its own sources.
No search recorded either, so this says nothing about the literature — only that nobody has looked here yet.
Code
No repository linked yet. Contribute one.
Sources
Whether a technique works is a claim. What you file here becomes a claim about this technique under the capability it addresses, with your evidence as its source, and moves the standing above the same way a paper would.