datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
RobloxAgent-toolreason-v1-2500!! We are not related to the Roblox company, this is a community dataset called RobloxAgent due to it being related to the Roblox studio environment.
RobloxAgent tool reason v1 dataset
Version 1 of RobloxAgent, an agentic dataset designed to train generalists LLM's alike to work better with the Roblox Studio Luau environment. All data was generated in a real, live, working Roblox studio place, you could also use this dataset to see how the AI agents dealt with the studio.
All… See the full description on the dataset page: https://huggingface.co/datasets/Pinkstack/RobloxAgent-toolreason-v1-2500.claude-sonnet-5-2500x
NOTE: this dataset does not have chain of thought, since anthropic doesn't provide it anymore
Information:
prompting dataset: https://huggingface.co/datasets/Roman1111111/prompts-for-claude-opus-4.6
distillation framework: https://github.com/caozhiyuan/copilot-api
Setts:
max tokens: 16384
reasoning: none