Sign in to view source links and access this dataset
Description
A synthetic supervised fine-tuning dataset containing 25,000 examples designed to mirror advanced reasoning, cybersecurity, coding, and agentic capabilities. It was created by WithinUsAI and converted to ShareGPT format by Alex01837178373. The dataset page was last updated on June 14, 2026.
Use Cases
Fine-tune language models for advanced reasoning tasks based on the described synthetic examples.
Train models on cybersecurity problem-solving scenarios based on the dataset's described focus.
Develop agentic AI capabilities based on the described synthetic training data.
Benchmark model performance on coding tasks based on the described dataset content.
Strengths
Contains 25,000 examples, providing a substantial volume for training.
Designed to emulate capabilities of frontier models like Anthropic's Claude Mythos/Fable.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
WithinUsAI/claude_mythos_distilled_25k
Collection Method
Synthetic generation
Freshness
Last updated 2026-06-14 21:51:40; freshness should be verified.
License is unknown; terms of use must be verified before application.