AINews news.smol.ai

Snowflake Arctic: Fully Open 10B+128x4B Dense-MoE Hybrid LLM

news.smol.ai·107w

OpenAI's Instruction Hierarchy for the LLM OS

news.smol.ai·107w

Perplexity, the newest AI unicorn

news.smol.ai·107w

FineWeb: 15T Tokens, 12 years of CommonCrawl (deduped and filtered, you're welcome)

news.smol.ai·107w

Llama-3-70b is GPT-4-level Open Model

news.smol.ai·108w

Meta Llama 3 (8B, 70B)

news.smol.ai·108w

Mixtral 8x22B Instruct sparks efficiency memes

news.smol.ai·108w

Lilian Weng on Video Diffusion

news.smol.ai·108w

Multi-modal, Multi-Aspect, Multi-Form-Factor AI

news.smol.ai·108w

Zero to GPT in 1 Year

news.smol.ai·109w

Mergestral, Meta MTIAv2, Cohere Rerank 3, Google Infini-Attention

news.smol.ai·109w

Music's Dall-E moment

news.smol.ai·109w

Gemini Pro and GPT4T Vision go GA on the same day by complete coincidence

news.smol.ai·109w

Anime pfp anon eclipses $10k A::B prompting challenge

news.smol.ai·109w

Mixture of Depths: Dynamically allocating compute in transformer-based language models

news.smol.ai·110w

Cohere Command R+, Anthropic Claude Tool Use, OpenAI Finetuning

news.smol.ai·110w

ReALM: Reference Resolution As Language Modeling

news.smol.ai·110w

Not much happened today

news.smol.ai·110w

AdamW -> AaronD?

news.smol.ai·110w

Evals-based AI Engineering

news.smol.ai·111w

Log in to enable infinite scrolling