datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
in-context-grid-reasoning
In-Context Grid Reasoning (ICGR)
A small, fully synthetic benchmark for demonstration-conditioned rule induction:
each task shows 2–4 (input grid → output grid) support pairs that share one
hidden transformation, and the model must apply the same transformation to a
held-out query input.
It targets the same behaviour probed by recent in-context / latent-reasoning work
on ARC-AGI (e.g. BDH-CQ: In-Context Learning with Recurrent Latent Reasoning,
arXiv:2608.09888), but is… See the full description on the dataset page: https://huggingface.co/datasets/WhySoCodius/in-context-grid-reasoning.CogSwitch-GridPuzzle-ReasoningSee the Git repo for more information.
Data format:Each file contains a list of question-reasoning pairs.E.g. from 341.json:
[
{
"id": 980,
"question": "The California condor ...",
"answer": "\n<recall>Clue #5 states ..."
},
...
]
There are roughly 5k question-answer pairs of increasing length between files. The reasoning chains were generated using GPT-4o-mini, and guided using the hints from the original GridPuzzle dataset. These chains have not been verified to be sound… See the full description on the dataset page: https://huggingface.co/datasets/Alex-Guha/CogSwitch-GridPuzzle-Reasoning.gridguard-instruct
GridGuard-Instruct Dataset
Overview
GridGuard-Instruct is a specialized dataset designed to train Large Language Models (LLMs) to act as autonomous Transmission System Operators (TSOs).
It contains 500+ synthetic examples of critical grid contingencies (Thermal Overloads, Voltage Collapses, Generator Trips) paired with correct, physics-compliant control actions in a structured JSON format.
Use Cases
Fine-tuning Agents: Train Llama-3 or Mistral to output valid… See the full description on the dataset page: https://huggingface.co/datasets/AnneshaChowdhury/gridguard-instruct.arc-agi-2-grids
ARC-AGI-2 Grids — training + analysis corpus (NVARC-compatible)
Companion dataset for the Kaggle ARC Prize 2026 (ARC-AGI-2) solver built on
sorokin/qwen3_4b_grids15_sft139 + per-task rank-256 LoRA (NVARC lineage).
Everything here is generated from public canonical data only (1,000
training / 120 evaluation tasks); no hidden competition data is included.
Contents
Path
Rows
Description
train/train_tasks.jsonl
1,000
canonical training tasks (full I/O)… See the full description on the dataset page: https://huggingface.co/datasets/Nabidnur/arc-agi-2-grids.