← Library

2026-01-09

5 saved articles

  2. Migrating our DOM to Zig

    Karl Seguin · Lightpanda

    We replaced LibDOM with a custom Zig implementation for better cohesion across events, Custom Elements, and ShadowDOM. Here's how we built it and what we learned along the way.

  5. Emergent temporal abstractions in autoregressive models enable hierarchical reinforcement learning

    [Submitted on 23 Dec 2025 (v1), last revised 24 Dec 2025 (this version, v2)] · arXiv.org

    Large-scale autoregressive models pretrained on next-token prediction and finetuned with reinforcement learning (RL) have achieved unprecedented success on many problem domains. During RL, these models explore by generating new outputs, one token at a time. However, sampling actions token-by-token can result in highly inefficient learning, particularly when rewards are sparse. Here, we show that it is possible to overcome this problem by acting and exploring within the internal representations o