# Data Sources, Licenses, and Attribution ## Overview This dataset was assembled by Domyn from publicly available datasets and from datasets generated by Domyn. We acknowledge the original dataset creators, contributors, and organizations responsible for collecting, curating, processing, and publishing the upstream data. The dataset incorporates material subject to different licenses and terms of use. The licensing information below describes the terms published by the corresponding upstream sources at the time of this release. These terms do not necessarily apply uniformly to every underlying record. The source references below are provided for attribution and provenance. They are not a representation that all underlying content has been independently cleared for redistribution. ## Source Attribution ### Conifer Source: ConiferLM Repository: https://huggingface.co/datasets/ConiferLM/Conifer Published license: Apache 2.0 ### Dolci-Instruct-SFT Source: Allen Institute for AI Repository: https://huggingface.co/datasets/allenai/Dolci-Instruct-SFT Published license: ODC-BY 1.0 ### Dolci-Instruct-SFT-Tool-Use Source: Allen Institute for AI Repository: https://huggingface.co/datasets/allenai/Dolci-Instruct-SFT-Tool-Use Published license: ODC-BY 1.0 Includes responses generated by GPT-4o, GPT-4.1 and GPT-5, subject to the OpenAI terms of use. ### Dolci-RL-Zero-IF-7B Source: Allen Institute for AI Repository: https://huggingface.co/datasets/allenai/Dolci-RL-Zero-IF-7B Published license: ODC-BY 1.0 ### Dolci-Think-SFT-7B Source: Allen Institute for AI Repository: https://huggingface.co/datasets/allenai/Dolci-Think-SFT-7B Published license: ODC-BY 1.0 ### Domyn IF Draft-Verify Source: Domyn Published license: generated by Domyn; responses by Qwen/Qwen3.6-27B (Apache 2.0), prompts from allenai/IF_multi_constraints_upto5 (ODC-BY 1.0) Each response drafts an answer, checks every constraint explicitly and revises; only responses that pass the rule-based checkers are kept. ### IF_multi_constraints_upto5_SFT Source: UniLu Repository: https://huggingface.co/datasets/UniLu/IF_multi_constraints_upto5_SFT Published license: ODC-BY 1.0 Responses generated by gemma-4-31B-it. ### IF_sft_data_verified Source: Allen Institute for AI Repository: https://huggingface.co/datasets/allenai/IF_sft_data_verified Published license: no license on the dataset card; released by Ai2 with the Dolci post-training data, whose new instruction-following data is licensed ODC-BY 1.0 ### ifeval-like-data Source: Argilla Repository: https://huggingface.co/datasets/argilla/ifeval-like-data Published license: Qwen License Responses generated by Qwen models. ### LMSYS-Chat-1M Source: LMSYS Repository: https://huggingface.co/datasets/lmsys/lmsys-chat-1m Published license: LMSYS-Chat-1M Dataset License Agreement The prompts are not redistributed. The rl_lmsys_redacted config gives the conversation id of each prompt; download them from the source after accepting its license. The instructions appended to each prompt are Domyn's. ### LoopTool-23k Source: zhangkangning Repository: https://huggingface.co/datasets/zhangkangning/LoopTool-23k Published license: Apache 2.0 ### Nemotron-Agentic-v1 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-Agentic-v1 Published license: CC BY 4.0 ### Nemotron-Post-Training-Dataset-v1 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-Post-Training-Dataset-v1 Published license: CC BY 4.0 Responses generated by Qwen3-235B-A22B. ### Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1 Published license: CC BY 4.0 ### Nemotron-RL-Agentic-Function-Calling-Pivot-v1 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1 Published license: CC BY 4.0 ### Nemotron-RL-Instruction-Following-Citation-Formatting-v1 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-Citation-Formatting-v1 Published license: CC BY 4.0 ### Nemotron-RL-Instruction-Following-Free-Form-Formatting-v1 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-Free-Form-Formatting-v1 Published license: CC BY 4.0 ### Nemotron-RL-Instruction-Following-Structured-Outputs-v2 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-RL-Instruction-Following-Structured-Outputs-v2 Published license: CC BY 4.0 Includes rows whose prompts Domyn edited to ask for the answer in a fenced code block. ### Nemotron-RL-instruction_following Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-RL-instruction_following Published license: ODC-BY 1.0 ### Nemotron-SFT-Agentic-v2 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-SFT-Agentic-v2 Published license: the card lists CC BY 4.0, Apache 2.0 and MIT; this release uses the tool_calling split ### Nemotron-SFT-Instruction-Following-Chat-v2 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-SFT-Instruction-Following-Chat-v2 Published license: ODC-BY 1.0 ### Nemotron-SFT-Instruction-Following-Chat-v3 Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/Nemotron-SFT-Instruction-Following-Chat-v3 Published license: the card lists ODC-BY, CC BY 4.0, Apache 2.0 and other; this release uses the instruction_following split, whose prompts are synthetic or seeded from allenai/tulu-3-sft-personas-instruction-following ### RECAST-30K Source: zk-guo Repository: https://huggingface.co/datasets/zk-guo/RECAST-30K Published license: CC BY 4.0 Only the responses written by deepseek-chat and qwen-plus are kept. ### SmolTalk2 Source: Hugging Face Repository: https://huggingface.co/datasets/HuggingFaceTB/smoltalk2 Published license: no license in the dataset card metadata; the card releases its new subsets, including multi-turn-reasoning-if (the one used here), under Apache 2.0 ### ToolACE Source: Team-ACE Repository: https://huggingface.co/datasets/Team-ACE/ToolACE Published license: Apache 2.0 ### ToolMind Source: Nanbeige Repository: https://huggingface.co/datasets/Nanbeige/ToolMind Published license: Apache 2.0 Rows regenerated from Salesforce/APIGen-MT-5k prompts (CC BY-NC 4.0) were removed. ### ToolRL Source: Cheng Qian et al. Repository: https://github.com/qiancheng0/ToolRL Published license: Apache 2.0 Built from ToolACE, Hammer and xLAM data. ### Toucan-1.5M Source: Agent-Ark Repository: https://huggingface.co/datasets/Agent-Ark/Toucan-1.5M Published license: Apache 2.0 ### When2Call Source: NVIDIA Repository: https://huggingface.co/datasets/nvidia/When2Call Published license: CC BY 4.0 ### xlam-function-calling-60k Source: Salesforce Repository: https://huggingface.co/datasets/Salesforce/xlam-function-calling-60k Published license: CC BY 4.0 ### xlam-irrelevance-7.5k Source: MadeAgents Repository: https://huggingface.co/datasets/MadeAgents/xlam-irrelevance-7.5k Published license: CC BY 4.0 ### xlam-irrelevance-7.5k (Qwen2.5-72B distill) Source: minpeter Repository: https://huggingface.co/datasets/minpeter/xlam-irrelevance-7.5k-qwen2.5-72b-distill-parsed Published license: CC BY 4.0; Qwen License (responses) Responses generated by Qwen2.5-72B-Instruct.