Niklas Muennighoff

@muennighoff

ai/llms @spacex @stanford

following
followers

PrimeIntellect is extending its multi-agent training infrastructure to swarms

PrimeIntellect says messages cost tokens and lose information, while agents must coordinate to avoid duplicating work.

15 Sources, 10d ago, updated 10d ago

A talk's influence on a PhD student's thesis conclusions and future work

A PhD student calls it the best talk they watched during their PhD.

4 Sources, 10d ago, updated 10d ago

Stanford CS25: V4 I Hyung Won Chung of OpenAI

YouTube

LLM context-removal methods may mostly hinge on how much of the prefix they keep, one post suggests

The author calls this a cleaner motivation for “prefix sliding” than the one in their own linked paper, “Prefix Sliding for efficient test-time scaling.”

1 Source, 43d ago, first seen 43d ago

LLMs Outperform Embedding Models at Much Higher Cost

Researcher shares findings comparing LLMs to specialized embedding models on benchmarks.

7 Sources, 61d ago, first seen 61d ago

Featured in

PrimeIntellect is extending its multi-agent training infrastructure to swarms

A talk's influence on a PhD student's thesis conclusions and future work

LLM context-removal methods may mostly hinge on how much of the prefix they keep, one post suggests

LLMs Outperform Embedding Models at Much Higher Cost

Featured in

PrimeIntellect is extending its multi-agent training infrastructure to swarms

A talk's influence on a PhD student's thesis conclusions and future work

LLM context-removal methods may mostly hinge on how much of the prefix they keep, one post suggests

LLMs Outperform Embedding Models at Much Higher Cost