goal612/Affine-UC-dpo-merged

TEXT GENERATIONConcurrent Unit Cost:3Model Size:35.1BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 21, 2026Architecture:Transformer Featherless Exclusive Cold

The goal612/Affine-UC-dpo-merged is a 35.1 billion parameter language model, derived from a LoRA-merged checkpoint of kevin954/Affine-5dfqbbh8ev-sft. This model is noted as a private TTL insurance project, indicating a specialized, potentially domain-specific application. Its development focuses on a specific, non-public use case, suggesting optimization for particular internal tasks rather than general-purpose applications.

Loading preview...

Overview

The goal612/Affine-UC-dpo-merged is a substantial 35.1 billion parameter language model. It was created by merging a LoRA checkpoint from kevin954/Affine-5dfqbbh8ev-sft. The model's README explicitly states its origin as a "private TTL insurance" project, indicating a highly specialized and potentially proprietary development context.

Key Characteristics

  • Parameter Count: 35.1 billion parameters, suggesting significant capacity for complex language understanding and generation tasks.
  • Origin: LoRA-merged from kevin954/Affine-5dfqbbh8ev-sft, implying a fine-tuning or adaptation process from a base model.
  • Context Length: Supports a context length of 32768 tokens, allowing it to process and generate longer sequences of text.
  • Purpose: Described as a "private TTL insurance" project, which points to a specific, likely internal or domain-specific application rather than a general-purpose public release.

Intended Use Cases

Given its description as a "private TTL insurance" project, this model is likely intended for:

  • Specialized Domain Tasks: Processing and generating content related to the insurance sector, particularly for private or internal applications.
  • Internal Development: Use within a specific organization for tasks that require a large, fine-tuned language model.

Limitations

As the model is noted as a "private TTL insurance" project and not a public submission until a "Stage-5 gate clears," its general applicability and public performance benchmarks are not available. It is not designed for broad, general-purpose LLM tasks.