datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
SHSP_Snake_TB_Women_NepQA
Nepali ShareGPT Health & Social Statistics Dataset
A Nepali-language, ShareGPT-formatted, single-turn instruction-following dataset of 453 question–answer pairs derived from official Nepali government health and demographic statistical publications. Every question is a fact-based question in Nepali (Devanagari script), and every answer is a short, fact-grounded response extracted directly from the source statistical tables — no hallucinated or generated numbers.
This README… See the full description on the dataset page: https://huggingface.co/datasets/sabin1234/SHSP_Snake_TB_Women_NepQA.trainsnake-sft-8x8
Snake SFT Dataset
A dataset of 2,000 Snake gameplay episodes with step-by-step reasoning and planning.
Dataset Statistics
Episodes: 2,000
Training examples: 48,429
Board size: 8x8
Average apples per episode: 24.2
Episodes with 25+ apples: 984 (49.2%)
Data Format
Each line contains a JSON object with:
task: Game identifier
answer: Episode summary (apples eaten, death status, steps)
reward: Number of apples eaten
prompt: System prompt for the AI
completion:… See the full description on the dataset page: https://huggingface.co/datasets/gokul8967/snake-sft-8x8.SnakeGPTsnake-linked-listssoeyi-themes