datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
linux-asm-pairs
Linux Kernel Assembly → Explanation Dataset
A dataset of disassembled Linux kernel functions paired with structured natural-language explanations, built from the Linux Commits Dataset (Zenodo 10654193).
Each row corresponds to a single compiled function extracted from a .c or .h file touched by a Bug-Fix Commit (BFC) or Bug-Introducing Commit (BIC) in the Linux kernel git history. Source files are compiled with gcc -O2 -g -fno-inline -fno-omit-frame-pointer and disassembled with… See the full description on the dataset page: https://huggingface.co/datasets/theelderemo/linux-asm-pairs.bringup_asm
BringUpBench C and Assembly
This dataset pairs C programs from BringUpBench 1.9 with assembly generated by Clang 17 for four CPU targets and seven optimization levels. It is intended for research on compilation, decompilation, assembly understanding, and cross-architecture code translation.
Dataset structure
The dataset contains 108 programs. Each optimization level is stored as a separate Hugging Face split, with 108 rows per split.
These are compiler… See the full description on the dataset page: https://huggingface.co/datasets/murodbek/bringup_asm.kakugo-asm
Kakugo Assamese dataset
[Paper] [Code] [Model]
A synthetically generated conversation dataset for training in Assamese.
This dataset contains synthetic conversational data and translated instructions designed to train Small Language Models (SLMs) for Assamese. It was generated using the Kakugo pipeline, a method for distilling high-quality capabilities from a large teacher model into low-resource language models. The teacher model used to generate this dataset was… See the full description on the dataset page: https://huggingface.co/datasets/ptrdvn/kakugo-asm.humaneval_asm
HumanEval-C and Assembly
This dataset pairs C functions from the HumanEval-Decompile benchmark with assembly generated by Clang 17 for four CPU targets and seven optimization levels. It is intended for research on compilation, decompilation, assembly understanding, and cross-architecture code translation.
Dataset structure
The dataset contains 164 programs. Each optimization level is stored as a separate Hugging Face split, with 164 rows per split.
These are… See the full description on the dataset page: https://huggingface.co/datasets/murodbek/humaneval_asm.mceval_asm
McEval-C and Assembly
This dataset pairs the C-language tasks from McEval with assembly generated by Clang 17 for four CPU targets and seven optimization levels. It is intended for research on compilation, decompilation, assembly understanding, and cross-architecture code translation.
Dataset structure
The dataset contains 50 programs. Each optimization level is stored as a separate Hugging Face split, with 50 rows per split.
These are compiler configurations, not… See the full description on the dataset page: https://huggingface.co/datasets/murodbek/mceval_asm.arm-asmasm-corpus
AsmCorpus — Assamese Pretraining Dataset
The largest open monolingual Assamese corpus for LLM pretraining.
Documents: 2.37M
Characters: 11B
GPT-2 tokens: ~3.7B | Gemma 4 E2B tokens: ~5.8B
Format: Parquet (text column only)
License: ODC-By 1.0
Usage
from datasets import load_dataset
ds = load_dataset("ananddey/asm-corpus", split="train", streaming=True)
for doc in ds:
print(doc["text"])
How It Was Built
All documents passed through language… See the full description on the dataset page: https://huggingface.co/datasets/ananddey/asm-corpus.x86-asm-instructions-23k
x86 ASM Instructions - 23K
Instruction-tuning dataset of x86 and x86_64 assembly snippets paired with short natural-language instructions or comments.
Dataset Summary
Field
Value
Total rows
23,104
Split(s)
train
Format
JSONL (instruction, output, system)
Instruction average length
47.0 characters
Instruction max length
413 characters
Output average length
514.25 characters
Output max length
15,805 characters
System prompt values
1… See the full description on the dataset page: https://huggingface.co/datasets/NickIBrody/x86-asm-instructions-23k.arm-asm-xsmalllegal-clause-instruction-Tunning
annotations_creators: [human-annotated, machine-generated]
language: [en]
license: apache-2.0
task_categories: [text-generation]
task_ids: [language-modeling]
pretty_name: Legal Clause Instruction Dataset
size_categories: 10K<samples<100K
🧾 Legal Clause Instruction Dataset
This dataset is designed to fine-tune large language models (LLMs) for structured legal document understanding — specifically clause identification, classification, and risk severity scoring. It is… See the full description on the dataset page: https://huggingface.co/datasets/asm3515/legal-clause-instruction-Tunning.okr-asms-corpus
OKR Agent Synthetic Corpus (ASMS)
A synthetic training corpus for an OKR (Objectives and Key Results) management agent, generated using Agent-Specific Model Synthesis (ASMS) — a pipeline that uses large LLMs as compilers to produce training data for task-specific micro-models.
Dataset Description
5,759 (input, tool_calls, methodology_notes) training triples covering 6 OKR management workflows against the Keyflow MCP API.
Each example maps a natural language user query to:… See the full description on the dataset page: https://huggingface.co/datasets/chan4lk/okr-asms-corpus.