messages stringlengths 200k 1.17M | model stringclasses 1
value | quality_scores stringclasses 6
values |
|---|---|---|
"[{\"role\": \"user\", \"content\": \"Build out a prototype for this potential system for stock AI a(...TRUNCATED) | claude-fable-5.5 | {"overall": 1.0, "source": "crownelius"} |
"[{\"role\": \"system\", \"content\": \"You are an autonomous coding agent running on the user's mac(...TRUNCATED) | claude-fable-5.5 | {"overall": 1.0, "source": "base_v1"} |
"[{\"role\": \"system\", \"content\": \"You are an autonomous coding agent running on the user's mac(...TRUNCATED) | claude-fable-5.5 | {"overall": 0.2, "source": "base_v1"} |
"[{\"role\": \"system\", \"content\": \"You are an autonomous coding agent running on the user's mac(...TRUNCATED) | claude-fable-5.5 | {"overall": 0.2, "source": "base_v1"} |
"[{\"role\": \"system\", \"content\": \"You are an autonomous coding agent running on the user's mac(...TRUNCATED) | claude-fable-5.5 | {"overall": 0.9, "source": "base_v1"} |
"[{\"role\": \"system\", \"content\": \"You are an autonomous coding agent running on the user's mac(...TRUNCATED) | claude-fable-5.5 | {"overall": 0.2, "source": "base_v1"} |
"[{\"role\": \"system\", \"content\": \"You are an autonomous coding agent running on the user's mac(...TRUNCATED) | claude-fable-5.5 | {"overall": 0.2, "source": "base_v1"} |
"[{\"role\": \"system\", \"content\": \"You are an autonomous coding agent running on the user's mac(...TRUNCATED) | claude-fable-5.5 | {"overall": 0.2, "source": "base_v1"} |
"[{\"role\": \"system\", \"content\": \"You are an autonomous coding agent running on the user's mac(...TRUNCATED) | claude-fable-5.5 | {"overall": 1.0, "source": "base_v1"} |
"[{\"role\": \"system\", \"content\": \"You are an autonomous coding agent running on the user's mac(...TRUNCATED) | claude-fable-5.5 | {"overall": 0.2, "source": "base_v1"} |
Dataset Description
This dataset contains 1,000 agentic coding and max reasoning multi-turn, high-quality traces generated by Fable 5.1 max recently (leaked version of fable 5.5).
It holds around 75,000,000 training-ready, reasoning-heavy tokens of step-by-step chain-of-thought programming across multiple complex domains.
It has also been deduplicated and heavily filtered to remove low-quality traces, keeping only high-quality traces.
Dataset Stats (FULL)
| Metric | Value |
|---|---|
| Total Examples | 1,000 Traces |
| Total Token Count | ~75,000,000 Tokens |
| Total Dataset Size | 284 MB |
| Average Trace Size | 284 KB |
| Average Token Count | ~75,000 Tokens |
Dataset Contents & Coverage
The dataset includes step-by-step problem-solving for complex coding tasks, including:
Algorithm design, implementation, and performance optimization.
Advanced agentic software engineering.
Multi-step logic design and compliance with complex prompt constraints.
Uses
Distilling Fable 5.1/Fable 5.5 agentic coding and reasoning capabilities down to smaller LLMs.
Improve general coding, tool-use and agentic reasoning quality.
Teaching models to generate clear chain-of-thought steps and tool-use before outputting their final answer.
- Downloads last month
- 59