datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
English-Zomi-OPUS_Tatoeba_v20230412
English–Zomi Parallel Corpus (1.78M)
This dataset contains 1.78 million English–Zomi sentence pairs, created to support
machine translation, linguistic research, and large‑scale language model training.
It is fully open and permissively licensed for commercial and non‑commercial use.
🌐 Linguistic Background: Zomi, Tedim Chin, and ISO Codes
Zomi is the endonym (self‑chosen name) of the people and their language.However, Zomi does not yet have an official ISO 639‑3 code.… See the full description on the dataset page: https://huggingface.co/datasets/ZomiLearner/English-Zomi-OPUS_Tatoeba_v20230412.Opus-WritingPrompts
Opus Writing Prompts
This is a dataset containing 3008 short stories, generated by an unrestrained Claude Opus using Reddit's Writing Prompts as a source. Each sample is generally between 4000-6000 characters long.
These stories were thoroughly cleaned and then further enriched with a title and a series of applicable genres.
Disclaimer: This dataset is extremely varied and includes erotica. You have been warned.
Three files are included:
A ShareGPT dataset, ready to be used for… See the full description on the dataset page: https://huggingface.co/datasets/Gryphe/Opus-WritingPrompts.ouroboros-osworld-verified-opus5
Ouroboros on OSWorld-Verified: 90.69%, the highest result reported to date
Status: Self-reported result over all 361 tasks. The official per-task
scores, prompts, manifests and feasibility records are public here, together
with every acting task record that the run produced.
Start here
Result
90.69% (327.39 / 361)
Model
anthropic/claude-opus-5
Method
Screenshot only, one rollout, 100 policy turns
Exact evidence
f52ebf2 and evidence.json… See the full description on the dataset page: https://huggingface.co/datasets/razzant/ouroboros-osworld-verified-opus5.claude-opus-4.6-4.7-reasoning-8.7k
Background
Ended up with some tokens to burn on a Claude Max plan. Assembly began during 4.6 and moved to 4.7. Model is tagged. The development evolved as it went along. The dataset has not been manually reviewed. It's entirely Claude developed.
Clarification on Reasoning
The reasoning is not Claude's actual chain-of-thought (cot) and is not summarized cot. It's a fully synthetic cot created as part of the Assistant response to mimic the type of "thinking"… See the full description on the dataset page: https://huggingface.co/datasets/angrygiraffe/claude-opus-4.6-4.7-reasoning-8.7k.kalo-opus-instruct-22k-no-refusalClaude-3-Opus-Instruct-15K
Original Character Card
Processed 15K Prompts - See Usable Responses Below
Based on Claude 3 Opus through AWS.
I took a random 5K + 10K prompt subset from Norquinal/claude_multi_instruct_30k to use as prompts, and called API for my answers.
Warning!
Uncleaned - Only Filtered for Blatant Refusals.
I will be going through and re-prompting missing prompts, but I do not expect much success, as some of the prompts shown are nonsensical, incomplete, or impossible… See the full description on the dataset page: https://huggingface.co/datasets/nothingiisreal/Claude-3-Opus-Instruct-15K.Opus_WritingStruct
Opus Writing Instruct 6k
Synthetically generated creative writing data using Claude 3 Opus, by Anthropic, filtered and cleaned using automated means. Focus was placed on having as many genres as possible represented in the data, and to have Claude more openly use its excellent prose.
It also contains question-answer instruction pairs related to the topic of writing.
Dataset Details
Curated by: Nopm
License: Apache 2
Credits: The entire SillyTilly community for providing… See the full description on the dataset page: https://huggingface.co/datasets/Nopm/Opus_WritingStruct.wildclaw-opus-traces
WildClaw Agent Traces — Claude Opus 4.6
Full agentic traces from running WildClawBench tasks through Claude Opus 4.6 via an instrumented reverse proxy.
Dataset Description
Each trace file captures the complete HTTP-level request/response pairs between the OpenClaw agent and Claude Opus 4.6, including:
System prompts, user messages, and assistant responses
Tool calls and tool results (multi-turn agentic loops)
Token usage and cost metadata from OpenRouter
Key… See the full description on the dataset page: https://huggingface.co/datasets/sammshen/wildclaw-opus-traces.Claude-opus-4.7-TraceInversion-5000x
🌀 Claude-opus-4.7-TraceInversion-5000x
v1.0 Release
A High-Fidelity Reconstructed CoT Dataset Saturated with the 'Opus Deep Logic Style' via Trace Inversion
📊 5,000 Samples
🧬 Trace Inversion & Negentropy
🛠 SFT & DPO Ready
🔥 Claude 4.7-Max Distillation
🌐 English & Multilingual
💡 What is Trace Inversion?
In Large Language Model (LLM) reasoning distillation, proprietary API models (such as GPT-4/5 and Claude)… See the full description on the dataset page: https://huggingface.co/datasets/Jackrong/Claude-opus-4.7-TraceInversion-5000x.claude-4.5-opus-high-reasoning-250xThis is a reasoning dataset created using Claude Opus 4.5 with a reasoning depth set to high. Some of these questions are from reedmayhew and the rest were generated.
The dataset is meant for creating distilled versions of Claude Opus 4.5 by fine-tuning already existing open-source LLMs.
Stats
Costs: $ 52.3 (USD)
Total tokens (input + output): 2.13 M
20260731_mini-v2.4.2_opus-5-xhighNOESIS-1M-reasoning-router-code-math-psych-opus47-deepseek4-qwen36-gemini31-r1-gpt54
⚡ Each donation funds the next large quant.
I host free GGUF or MoE quants as independent research.
Local hardware: Mechrevo Kuangshi GM7AG0M — RTX 3060 Laptop 6GB GDDR6, 64GB DDR5, i7-12700H (14C/20T, 4.7GHz), Windows 11, Samsung 990 Pro.
Good for imatrix and 0.6–35B-class work in RAM. 9B+ and searches need rented H200/Blackwell, typically $100 per quant.
🎉 Boosty🦄 |
☕ Buy Me a Coffee🦄 |
⭐ DonationAlerts🦄
💚 Thanks to Hugging Face for extra storage.🦄… See the full description on the dataset page: https://huggingface.co/datasets/AMAImedia/NOESIS-1M-reasoning-router-code-math-psych-opus47-deepseek4-qwen36-gemini31-r1-gpt54.Opus-4.6-Reasoning-3000x-filtered
[!WARNING] NOTICE: The original dataset has been updated with better filtering. Please use the original dataset, not this one.
Filtered from: https://huggingface.co/datasets/crownelius/Opus-4.6-Reasoning-3000x
The original dataset has 979 refusals, I removed these in this version.
lordx64-claude-opus-4.7-max-cleaned
reasoning-distill-claude-opus-4-7-max-cleaned
Cleaned version of lordx64/reasoning-distill-claude-opus-4-7-max.
See the original dataset for full provenance, collection methodology, and terms of use.
Cleaning steps
Step
Filter
Reason
Rows removed
1
Simulated thinking (...)
Rows with ... in thinking/response indicate the model learned to simulate reasoning (e.g., "Now I'm laying out the puzzle grids...") rather than actually performing it. This causes failures… See the full description on the dataset page: https://huggingface.co/datasets/TeichAI/lordx64-claude-opus-4.7-max-cleaned.Opus_Instruct_25kA scaled-up continuation of what I did with Opus Instruct 3k.
Quality checks were not very thorough so far, I did some general keyword filtering and parsing to make sure that there wasn't turn leakage.
This time I made sure most responses were two turns.
NOESIS-50K-reasoning-router-code-math-psych-opus47-deepseek4-qwen36-gemini31-r1-gpt54
⚡ Each donation funds the next large quant.
I host free GGUF or MoE quants as independent research.
Local hardware: Mechrevo Kuangshi GM7AG0M — RTX 3060 Laptop 6GB GDDR6, 64GB DDR5, i7-12700H (14C/20T, 4.7GHz), Windows 11, Samsung 990 Pro.
Good for imatrix and 0.6–35B-class work in RAM. 9B+ and searches need rented H200/Blackwell, typically $100 per quant.
🎉 Boosty🦄 |
☕ Buy Me a Coffee🦄 |
⭐ DonationAlerts🦄
💚 Thanks to Hugging Face for extra storage.🦄… See the full description on the dataset page: https://huggingface.co/datasets/AMAImedia/NOESIS-50K-reasoning-router-code-math-psych-opus47-deepseek4-qwen36-gemini31-r1-gpt54.opus-gpt-swe-frontier-core
SWE Base
Repository-level software engineering trajectories for training coding agents.
2,459 chat trajectories · 48,499 API calls · $837.57 recorded generation cost
SWE-bench · debugging · patching · tools · agents
Overview
SWE Base is a software-engineering dataset centered on real repository issues. Each training example gives an agent a problem statement and captures the multi-turn process of inspecting a codebase, reasoning about a bug… See the full description on the dataset page: https://huggingface.co/datasets/Roman1111111/opus-gpt-swe-frontier-core.20260507_mini-v2.2.6_opus-4-7-xhighclaude_opus_4.8_max_thinking_5k_v2
Claude Opus 4.8 MAX THINKING — Distillation Dataset
5,000 high-quality examples designed to distill the maximum-effort reasoning, honest analysis, production software engineering, and agentic capabilities of Claude Opus 4.8.
Overview
This dataset captures Opus 4.8’s signature strengths:
Deep, structured, high-effort reasoning
Honest communication about trade-offs and uncertainties
Excellent production software engineering judgment
Strong agentic workflow design… See the full description on the dataset page: https://huggingface.co/datasets/11-47/claude_opus_4.8_max_thinking_5k_v2.Claude-opus-4.6-TraceInversion-9000x
🌀 Claude-opus-4.6-TraceInversion-9000x
v1.0 Release
A High-Fidelity Reconstructed CoT Dataset via Trace Inversion
📊 9,000 Samples
🧬 Trace Inversion & Negentropy
🛠 SFT & DPO Ready
🔥 Claude 4.6 Distillation
🌐 English & Multilingual
💡 What is Trace Inversion?
In Large Language Model (LLM) reasoning distillation, proprietary API models (such as GPT-4/5 and Claude) typically hide their internal thinking steps, providing… See the full description on the dataset page: https://huggingface.co/datasets/Jackrong/Claude-opus-4.6-TraceInversion-9000x.claude-opus-4.6-10000xThis is a high-fidelity reasoning dataset synthesized using Claude Opus 4.6. The dataset is designed to capture the model's internal "Chain of Thought" and reasoning traces, specifically focusing on mathematical accuracy and structured logical deduction.
The dataset is intended for Supervised Fine-Tuning (SFT) and Distillation, allowing smaller open-source models to inherit the sophisticated reasoning patterns of Claude Opus 4.6.
Dataset Description
This collection combines high-difficulty… See the full description on the dataset page: https://huggingface.co/datasets/Roman1111111/claude-opus-4.6-10000x.swe-bench-opus-logs
Claude 3 inference SWE-Bench results
Contains prompting responses from SWE-bench on these 2 settings:
Oracle retrieval
BM25 retrieval
Each of the subsets contains an additional log_last_line attributes which is the last line from log files generated during evaluation step.
Results:
Model
BM25 Retrieval Resolved (%)
Oracle Retrieval Resolved (%)
GPT-4*
0
1.74
Claude-2
1.96
4.80
Claude-3 Opus (20240229)
3.24
6.42
Claude-2 and GPT-4 results from SWE-bench… See the full description on the dataset page: https://huggingface.co/datasets/syntaxsynth/swe-bench-opus-logs.claude-opus-4.8-pi-tracesMore expensive than anticpated so you only get 4 lol :P
This dataset was generated using teich by TeichAI
Prepare these datasets for supervised fine-tuning in just a few lines of code — see the Conversion section below.
Claude Opus 4.8 Pi Traces
This directory contains raw agent trace files generated by teich.
All assistant responses were generated by anthropic/claude-opus-4.8.
JSONL files: 4
Training-ready tools
A complete configured tools schema snapshot is… See the full description on the dataset page: https://huggingface.co/datasets/Quaxicron/claude-opus-4.8-pi-traces.badlogicgames-pi-mono-opus-filteredFiltered version of badlogicgames/pi-mono - Only opus traces, dropped invalid sessions as well.
All traces present are training safe and teich compatible
20260429_mini-v2.2.6_opus-4-7Dans-Prosemaxx-Opus-Writingclaude-opus-4.8-pi-tracesMore expensive than anticpated so you only get 4 lol :P
This dataset was generated using teich by TeichAI
Prepare these datasets for supervised fine-tuning in just a few lines of code — see the Conversion section below.
Claude Opus 4.8 Pi Traces
This directory contains raw agent trace files generated by teich.
All assistant responses were generated by anthropic/claude-opus-4.8.
JSONL files: 4
Training-ready tools
A complete configured tools schema snapshot is… See the full description on the dataset page: https://huggingface.co/datasets/armand0e/claude-opus-4.8-pi-traces.Opus_Instruct_3k~2.5 million tokens worth of general purpose multi-turn instruction finetuning data in the style of Claude 3 Opus.
Turns for both the Human and Assistant are created by the model itself (ala Magpie).
opus-agent-merged-v2claude-opus-4.6-4.7-reasoning-8.7k
Background
Ended up with some tokens to burn on a Claude Max plan. Assembly began during 4.6 and moved to 4.7. Model is tagged. The development evolved as it went along. The dataset has not been manually reviewed. It's entirely Claude developed.
Clarification on Reasoning
The reasoning is not Claude's actual chain-of-thought (cot) and is not summarized cot. It's a fully synthetic cot created as part of the Assistant response to mimic the type of "thinking" expected to… See the full description on the dataset page: https://huggingface.co/datasets/Mahfug/claude-opus-4.6-4.7-reasoning-8.7k.