unconst/Affine-5czsc2fc98-r499-sbsv5-offline-dpo-hialpha-midrank-lobeta-softctx-extrasteps-merged

TEXT GENERATIONPricing:Input $0.4 / Cached $0.07 / Output $4Concurrent Unit Cost:3Model Size:35.1BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 16, 2026Architecture:Transformer Featherless Exclusive Cold

The unconst/Affine-5czsc2fc98-r499-sbsv5-offline-dpo-hialpha-midrank-lobeta-softctx-extrasteps-merged model is a 35.1 billion parameter language model, merged from a LoRA fine-tuned version of kevin954/Affine-5dfqbbh8ev-sft. With a context length of 32768 tokens, this model is designed for specific, private applications, indicated by its "private TTL insurance" status. Its primary differentiator lies in its specialized merging process, suggesting an optimization for particular downstream tasks rather than general-purpose use.

Loading preview...

Model Overview

The unconst/Affine-5czsc2fc98-r499-sbsv5-offline-dpo-hialpha-midrank-lobeta-softctx-extrasteps-merged is a substantial language model featuring 35.1 billion parameters and a 32768-token context length. This model is a merged checkpoint, specifically derived from a LoRA fine-tuned version of kevin954/Affine-5dfqbbh8ev-sft.

Key Characteristics

  • Parameter Count: 35.1 billion, indicating a large-scale model capable of complex language understanding and generation.
  • Context Length: A generous 32768 tokens, allowing for processing and generating extensive texts while maintaining coherence.
  • Origin: It is a merged checkpoint, suggesting a combination of different training stages or fine-tuning approaches to achieve its current state.
  • Development Status: Described as having "private TTL insurance" and not a public submission, implying it is intended for specific, internal, or proprietary applications.

Intended Use Cases

Given its specialized merging process and private status, this model is likely tailored for:

  • Specific Downstream Applications: Optimized for particular tasks or domains where the base model kevin954/Affine-5dfqbbh8ev-sft was fine-tuned.
  • Proprietary Research and Development: Its private nature suggests use within a controlled environment for specialized problem-solving or experimentation.

This model is not presented as a general-purpose LLM but rather as a highly customized solution for targeted applications.

Finetunes
1 models