unconst/Affine-5czsc2fc98-r384-offline-dpo-hialpha-hirank-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-r384-offline-dpo-hialpha-hirank-merged model is a 35.1 billion parameter language model, derived from a LoRA merge of kevin954/Affine-5dfqbbh8ev-sft. With a context length of 32768 tokens, this model is noted as a private TTL insurance checkpoint. Its primary characteristic is being a merged checkpoint salvage, indicating a focus on specific, potentially internal, applications rather than general-purpose use.

Loading preview...

Model Overview

The unconst/Affine-5czsc2fc98-r384-offline-dpo-hialpha-hirank-merged model is a substantial 35.1 billion parameter language model with a context length of 32768 tokens. It is specifically identified as a LoRA-merged checkpoint, originating from kevin954/Affine-5dfqbbh8ev-sft.

Key Characteristics

  • Parameter Count: 35.1 billion parameters, indicating a large-scale model capable of complex language understanding and generation tasks.
  • Context Length: Supports a significant context window of 32768 tokens, allowing it to process and generate longer sequences of text while maintaining coherence.
  • Origin: This model is a result of a LoRA (Low-Rank Adaptation) merge from a base model, kevin954/Affine-5dfqbbh8ev-sft.
  • Purpose: Described as a "private TTL insurance" and a "merged checkpoint salvage," suggesting it serves a specialized, potentially internal or recovery-oriented function rather than being a general-purpose public release.

Intended Use Cases

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

  • Specialized Internal Applications: Potentially used within a specific organization for tasks requiring a robust, large-scale language model.
  • Research and Development: Could be a checkpoint for ongoing research, fine-tuning, or evaluation within a controlled environment.
  • Data Recovery or Archival: The term "salvage" might imply its role in preserving or recovering model states for future use or analysis.