British startup turns video game inputs into training data for physical-world AI
WIRED reports that a British startup is converting video game inputs into training data for AI models designed to navigate the physical world.
#training-data
View all AI newsAi2 says Marin’s latest hero run includes about 1.8T Dolma 3.5 tokens
Ai2 says Marin’s latest hero run includes about 1.8 trillion tokens from its Dolma 3.5 corpus, and that Marin’s data-mixture work draws on Ai2’s Olmix recipes and Organize the Web curation methodology. Will Held says Marin’s 535B run was built on 25 trillion tokens from 152 datasets licensed for training.
Space Bunny Alpha is free, and its provider does not train on prompts or completions
OpenRouter says the stealth model Space Bunny Alpha is free. This time, the provider does not train on prompts or completions.