Crossposted from the AI Alignment Forum. May contain more technical jargon than usual.This is a linkpost for https://github.com/ucl-dark/llm_debate/blob/main/paper.pdf

We've just completed a bunch of empirical work on LLM debate, and we're excited to share the results. If the title of this post is at all interesting to you, we recommend heading straight to the paper. There are a lot of interesting results that are hard to summarize, and we think the paper is quite readable.

If you're pressed for time, we've posted the abstract and our Twitter thread below.

If you're working on debate or might in future, we especially suggest reading our recommendations for working on debate (below or in Appendix C of the paper).

Code: https://github.com/ucl-dark/llm_debate
Examples: https://llm-debate.com
Paper: https://github.com/ucl-dark/llm_debate/blob/main/paper.pdf

Abstract.

Common methods for aligning large language models (LLMs) with desired behaviour heavily rely on human-labelled data. However, as models grow increasingly sophisticated, they [...]

---

Outline:

(02:13) Twitter thread

(05:36) Further LLM judge results

(06:56) Further human judge results

(08:56) Recommendations for working on debate

(09:00) Dataset recommendations

(11:31) Iteration recommendations

(13:39) Implementation recommendations

---

First published:
February 7th, 2024

Source:
https://www.lesswrong.com/posts/2ccpY2iBY57JNKdsP/debating-with-more-persuasive-llms-leads-to-more-truthful

Linkpost URL:
https://github.com/ucl-dark/llm_debate/blob/main/paper.pdf

---

Narrated by TYPE III AUDIO.

[Linkpost] “Debating with More Persuasive LLMs Leads to More Truthful Answers” by Akbir Khan, John Hughes, Dan Valentine, Sam Bowman, Ethan Perez

LessWrong (30+ Karma)