https://hackerpaper.kaipfister.de/p/2312.06681/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMThttps://news.ycombinator.com/item?id=41295833urn:hackerpapers:mention:2312.06681:41295833Tue, 20 Aug 2024 00:58:16 GMT<p>Ty! - We generate contrast pairs (for this post, using gpt-4o) and do some post processing for quality (synthetic data). The impact of different types of contrast pairs is a continuing area of research for us. - We treat the evaluator model layers as hyper parameters, similar to the steering research (some of which we cite in our “non-comprehensive list of references”). We also see that the middle layers tend to be most effective. - For base model, we use both Llama-3-8b-Instruct and Llama-3.1-8b-Instruct to show LSR taking advantage of the improved base model (maybe I misunderstood the…</p><p><a href="https://hackerpaper.kaipfister.de/p/2312.06681/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.CL