‹ BackHN Continuity

Thread

Clef: Open-weight decision models, and new RL fine-tuning platform

637 points · 217 comments · jasondavies

  1. iugtmkbdfil834 · · focus · HN ↗
    I am playing with it now trying to make it play adom ( mostly to avoid issues with timing; using local variants so speed is a consideration ). By itself, it plays in ways you can likely imagine.. not well. Adding big Qwen on top for strategic decisions helps, but makes it fail in unexpected ways. Will be adding smaller non-qwen in the middle. WIP so actual details are changing fast.
    1. ramoz · · focus · HN ↗
      Idk what adom is, but somewhat related I had jev play poker and it was abysmal. <a href="https:&#x2F;&#x2F;backnotprop.com&#x2F;blog&#x2F;jev-poker&#x2F;" rel="nofollow">https:&#x2F;&#x2F;backnotprop.com&#x2F;blog&#x2F;jev-poker&#x2F;
      1. iugtmkbdfil834 · · focus · HN ↗
        <a href="https:&#x2F;&#x2F;www.adom.de&#x2F;home&#x2F;index.html" rel="nofollow">https:&#x2F;&#x2F;www.adom.de&#x2F;home&#x2F;index.html

        Sorry, sometimes I forget not everyone is me;p

        Hmm,where do you think jev could fit?

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.