Count Cpt Corpus V1: Text for Voice and Doctrine Fine-Tuning

A Continued Pretraining (CPT) corpus for fine-tuning The Count, a voice-and-doctrine fine-tune project. The corpus was created by Sequoyah Kennedy and collaborators, with a major content expansion written on 2026-05-21. The dataset includes 17 diagnostic essays and a 14-chapter work.

Use Cases

• Fine-tuning language models for specific stylistic 'voice' based on the described content expansion.
• Continued pretraining of models on a corpus of diagnostic essays and chapters.
• Evaluating model performance on text that follows a particular doctrine or framework.
• Training models to generate text in the style of the provided essays and chapters.

Strengths

• Includes a major content expansion written on 2026-05-21.
• Contains 17 diagnostic essays and a 14-chapter work.
• Associated with a public GitHub repository containing harvesting and training scripts.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and license are unknown.
• Data may reflect the specific stylistic and doctrinal bias of its creators.

Provenance

Source

shkennedy33/count-cpt-corpus-v1 on GitHub.

Collection Method

Likely gathered and curated for the specific fine-tune project.

Time Range

Content expansion dated 2026-05-21.

Freshness

Last updated 2026-05-22 04:53:25.