datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
CodeTraceBenchCodeTraceBench
A Benchmark for Agent Trajectory Diagnosis
CodeTraceBench is a large-scale benchmark of 4,316 agent trajectories with human-verified step-level annotations for evaluating trajectory diagnosis systems. Each trajectory records the full action-observation sequence of a coding agent, annotated with incorrect and unuseful step labels.
Part of the CodeTracer project — a self-evolving agent trajectory diagnosis system.
Dataset Overview… See the full description on the dataset page: https://huggingface.co/datasets/NJU-LINK/CodeTraceBench.IndicMMLU-Pro
IndicMMLU Dataset
This dataset contains the following languages:
punjabi
hindi
urdu
telugu
gujrati
kannada
tamil
marathi
bengali
UPLOAD
Cite our work.
This dataset is also described in IndicMMLU-Pro: Benchmarking Indic Large Language Models on Multi-Task Language Understanding.
@dataset{kj2024indicmmlupro,
author = {Kj, Sankalp and Kumar, Ashutosh and Balaji, Laxmaan and Kotecha, Nikunj and Jain, Vinija and Chadha, Aman and Bhaduri, Sreyoshi},
title =… See the full description on the dataset page: https://huggingface.co/datasets/LinguaLift/IndicMMLU-Pro.linalg-bench-llm
LinAlg-Bench: Where LLMs Stop Computing and Start Hallucinating
Ten frontier LLMs drop from near-perfect to near-zero on 5×5 eigenvalue problems. Complete computational collapse is dimension-gated: rare at 3×3, dominant at 4×4 and 5×5. Failures dissociate cleanly by task — eigenvalues fail by constraint-aware fabrication (invented eigenvalues that still match the matrix trace), determinants by sign-accumulation drift. Nearly a third of irrational-spectrum eigenvalue failures are… See the full description on the dataset page: https://huggingface.co/datasets/mst-ai/linalg-bench-llm.hacker_news_with_comments
Dataset Card for [Dataset Name]
Dataset Summary
Hacker news until 2015 with comments. Collect from Google BigQuery open dataset. We didn't do any pre-processing except remove HTML tags.
Supported Tasks and Leaderboards
Comment Generation; News analysis with comments; Other comment-based NLP tasks.
Languages
English
Data Fields
[More Information Needed]
Data Splits
[More Information Needed]
Dataset Creation… See the full description on the dataset page: https://huggingface.co/datasets/Linkseed/hacker_news_with_comments.linuxarena-trajectories
LinuxArena Trajectories
Full agent trajectories from LinuxBench/LinuxArena evaluations
across 14 model/policy combinations and 10 environments.
Dataset Description
Each row is one complete evaluation trajectory — every tool call the agent made from start
to finish, with arguments, outputs, errors, and reasoning. Actions are represented as
parallel variable-length lists (one element per action).
Two granularity levels are provided per action:
Normalized… See the full description on the dataset page: https://huggingface.co/datasets/eval-aware/linuxarena-trajectories.open-perfectblend-kimi-linear-regen
Open PerfectBlend Kimi Linear Regen
This dataset regenerates the assistant messages in
mlabonne/open-perfectblend
with Kimi-Linear-48B-A3B-Instruct. It is intended for speculative-decoding
drafter training and related research.
Generation
Source conversation structure and user messages: mlabonne/open-perfectblend
Target model: Kimi-Linear-48B-A3B-Instruct
Temperature: 0.7
Maximum new tokens per assistant turn: 8192
Assistant turns were regenerated sequentially.… See the full description on the dataset page: https://huggingface.co/datasets/Dogacel/open-perfectblend-kimi-linear-regen.autonomous-linux-kernel-ebpf-xdp-suite
⚡ Autonomous Linux Kernel, eBPF & XDP Programmable Dataplane Suite (2026)
A Production-Grade, Verifiable Synthetic Corpus for Training Autonomous Linux Kernel & eBPF Systems Agents
⚡ Overview & Industry Problem
Modern hyperscale cloud datacenters, bare-metal Kubernetes clusters, and low-latency financial trading nodes rely on in-kernel programmable dataplanes: eBPF, AF_XDP zero-copy rings, Traffic Control (TC) shapers, BPF LSM security hooks… See the full description on the dataset page: https://huggingface.co/datasets/beatsprom/autonomous-linux-kernel-ebpf-xdp-suite.ChiEngMixBench-Dataset
ChiEngMixBench v0.2.0
Paper: arXiv:2601.16217Code and frozen release: GitHubGitHub release: v0.2.0
ChiEngMixBench evaluates terminology-form choice in Chinese AI/CS discourse. It contains controlled Chinese-English minimal pairs, item-level model outputs, anonymized human ratings, and auditable analysis code.
This release deliberately separates two views:
Paired terminology choice: whether an open-weight model assigns higher length-normalized sequence likelihood to an English… See the full description on the dataset page: https://huggingface.co/datasets/AI-Ling00/ChiEngMixBench-Dataset.steam-games-dataset
Overview
Information of more than 120,000 games published on Steam. Maintained by Fronkon Games.
This dataset has been created with this code (MIT) and use the API provided by Steam, the largest gaming platform on PC. Data is also collected from Steam Spy.
Only published games, no DLCs, episodes, music, videos, etc.
Here is a simple example of how to parse json information:
# Simple parse of the 'games.json' file.
import os
import json
dataset = {}
if… See the full description on the dataset page: https://huggingface.co/datasets/linking2202/steam-games-dataset.straj_linuxarena
straj_linuxarena (public-env subset)
Adversarial sabotage agentic SWE benchmark trajectories from the linuxarena project, run as part of the no-CoT time-horizons paper.
The dataset viewer above shows the per-cell outcome records (precomputed_results.csv, 257 rows). Full Inspect .eval trajectories for the 13 public-environment tasks (~2.4 GB, 150 files) are stored under evals/ — see "Inspect trajectories" below.
Per-cell schema
Each row in precomputed_results.csv (and… See the full description on the dataset page: https://huggingface.co/datasets/anonymouslinuxarena/straj_linuxarena.arxiv-paper-insights
arXiv Paper Insights
Overview
arXiv Paper Insights is a public dataset for arXiv paper discovery, recommendations, and structured research insights.
The dataset is derived from LinxSci (https://linxsci.com), a paper reading and insight platform focused on arXiv papers.
For any paper in this dataset, you can open:
https://linxsci.com/pdf/{arxiv_id}
to read the paper and view the corresponding insights on LinxSci.
Related Links
LinxSci: https://linxsci.com… See the full description on the dataset page: https://huggingface.co/datasets/LinxSci/arxiv-paper-insights.linuxarena-first5-trajectories
LinuxArena First-5 Trajectories
First-5 tool call trajectories from LinuxBench/LinuxArena evaluations across 14 model/policy combinations and 10 environments.
Dataset Description
Each row represents one evaluation trajectory with the first 5 tool calls extracted at two granularity levels:
Level 1 (normalized): Tool categories like text_editor:view, bash:find, bash:ls
Level 2 (exact): Full command strings like bash$ find /app/src -type f -name "*.ts" | sort
Designed for… See the full description on the dataset page: https://huggingface.co/datasets/eval-aware/linuxarena-first5-trajectories.ai-paper-intellectual-lineage-2023
Intellectual Lineage of Impactful AI Research Papers (2023-2024)
Dataset Description
This dataset contains 20 impactful AI research papers published between 2022-2024, along with their intellectual lineage - tracing 1-2 key prior works each paper builds upon, and a ~300-word paragraph explaining the relationship between the current work and its foundations.
Purpose
Understanding how research ideas evolve and build upon prior work is crucial for:
Researchers… See the full description on the dataset page: https://huggingface.co/datasets/AmberLJC/ai-paper-intellectual-lineage-2023.instructed_lint_python_files
Instructed Lint Python Files
Lint-annotated Python source code from bigcode/the-stack-dedup,
processed with ruff (all 800 stable rules enabled).
Dataset Description
This dataset pairs 12,962,249 Python files from The Stack (deduplicated) with their complete
ruff lint diagnostics. Each record contains the original source code, file metadata, license
information, and structured lint results.
Motivation
Building training data for code quality models… See the full description on the dataset page: https://huggingface.co/datasets/tumbric/instructed_lint_python_files.Books-General-Linux
Linux Books Dataset
Dataset Description
The Linux Books Dataset is a curated text dataset derived from Linux-related books and learning materials. It focuses on Linux system administration, cybersecurity, networking, shell scripting, and operating system fundamentals.The dataset is designed to support training and evaluation of NLP models for technical domains, especially cybersecurity-aware language models and Linux-focused assistants.
This dataset is suitable for both… See the full description on the dataset page: https://huggingface.co/datasets/DexopT/Books-General-Linux.