Researchers baked 3x inference speedups directly into LLM weights — without speculative decoding

Researchers from the University of Maryland, Lawrence Livermore, Columbia and TogetherAI have developed a training technique that triples LLM inference speed without auxiliary models or infrastructure ...
GitHub

slashrebootofficial/simulated-metacognition-in-open-source-llms

This repository archives artifacts (prompts, configs, logs, and scripts) from a series of preprints (more info at https://slashreboot.com) on prompt-induced simulated metacognition and embodiment in ...