Open Source AI

concepts · 9 notes linked

Related: Large Language Models · Openai · Meta · Stanford · Instruction Fine Tuning · Eric Schmidt · Kyutai · Mistral AI

Notes

• $450 and 19 hours is all it takes to rival OpenAI's o1-preview — UC Berkeley Sky-T1-32B open-source reasoning model built for $450 in 19 hours
• Billionaires including Eric Schmidt plow $300 million into a non-profit that is France's latest push to catch up in AI | Fortune — France's €300M open-source AI research nonprofit kyutai launched
• GitHub - tatsu-lab/stanford_alpaca: Code and documentation to train Stanford's Alpaca models, and generate the data. — Instruction-following LLaMA model fine-tuned via self-instruct data
• How the Foundation Model Transparency Index Distorts Transparency — EleutherAI critique of Stanford FMTI biasing transparency toward corporate services
• Open-Source AI Is Uniquely Dangerous — Unsecured open-source AI uniquely risky because safety features cannot be re-patched
• Open-source AI matches top proprietary model in solving tough medical cases — Open-source Llama 3.1 405B matches GPT-4 on clinical diagnostic reasoning
• Technology Innovation Institute Open-Sourced Falcon LLMs: A New AI Model That Uses Only 75 Percent of GPT-3's Training Compute, 40 Percent of Chinchilla's, and 80 Percent of PaLM-62B's — TII open-sources Falcon-40B and Falcon-7B compute-efficient decoder LLMs
• Timeline of Open and Proprietary Large Language Models | NextBigFuture.com — Chronological overview of open and proprietary LLM releases through 2023
• Why LLaMa Is A Big Deal — LLaMA enables GPT-3-class LLM inference on consumer hardware