This dataset contains 10,000 agentic coding and reasoning multi-turn traces generated by the new Fable 5.1 model using max reasoning effort.
It holds almost 500,000,000 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.
| Metric | Value |
|---|---|
| Total Examples | 10,000 Traces |
| Total Token Count | ~500,000,000 Tokens |
| Total Dataset Size | 1.97 GB |
| Average Trace Size | 197 KB |
| Average Token Count | ~50,000 Tokens |
The dataset includes step-by-step problem-solving for complex coding tasks, including:
Algorithm design, implementation, and performance optimization.
Advanced debugging and error-handling.
Multi-step logic design and compliance with complex prompt constraints.
Distilling Fable 5.1 agentic coding and reasoning capabilities down to smaller LLMs.
Improve general coding and reasoning quality.
Teaching models to generate clear chain-of-thought steps and tool-use before outputting their final answer.
42 commits
This dataset contains 10,000 agentic coding and reasoning multi-turn traces generated by the new Fable 5.1 model using max reasoning effort.
It holds almost 500,000,000 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.
| Metric | Value |
|---|---|
| Total Examples | 10,000 Traces |
| Total Token Count | ~500,000,000 Tokens |
| Total Dataset Size | 1.97 GB |
| Average Trace Size | 197 KB |
| Average Token Count | ~50,000 Tokens |
The dataset includes step-by-step problem-solving for complex coding tasks, including:
Algorithm design, implementation, and performance optimization.
Advanced debugging and error-handling.
Multi-step logic design and compliance with complex prompt constraints.
Distilling Fable 5.1 agentic coding and reasoning capabilities down to smaller LLMs.
Improve general coding and reasoning quality.
Teaching models to generate clear chain-of-thought steps and tool-use before outputting their final answer.
42 commits