cs-552-2026-MMRF/safety_model

TEXT GENERATIONConcurrency Cost:1Model Size:2BQuant:BF16Ctx Length:32kTool Calling:SupportedPublished:Jun 3, 2026Architecture:Transformer Cold

The cs-552-2026-MMRF/safety_model is a 2 billion parameter language model, fine-tuned from cs-552-2026-MMRF/15kDPO, with a 32768 token context length. This model is specifically optimized for safety-related applications, demonstrating a final validation loss of 0.0466. Its primary strength lies in its fine-tuned performance for specific safety tasks, making it suitable for content moderation and risk assessment.

Loading preview...

Model Overview

The cs-552-2026-MMRF/safety_model is a 2 billion parameter language model, fine-tuned from the cs-552-2026-MMRF/15kDPO base model. It features a substantial context length of 32768 tokens, enabling it to process and understand extensive inputs. The model was trained for 1 epoch with a learning rate of 2e-05 and a total batch size of 16, achieving a final validation loss of 0.0466.

Key Characteristics

  • Base Model: Fine-tuned from cs-552-2026-MMRF/15kDPO.
  • Parameter Count: 2 billion parameters.
  • Context Length: Supports up to 32768 tokens.
  • Performance: Achieved a validation loss of 0.0466, indicating strong performance on its evaluation set.
  • Training: Utilized AdamW_Torch_Fused optimizer and a linear learning rate scheduler.

Potential Use Cases

While specific intended uses are not detailed in the README, the model's name and fine-tuning suggest an orientation towards safety-critical applications. Developers might consider this model for:

  • Content Moderation: Identifying and flagging inappropriate or harmful content.
  • Risk Assessment: Analyzing text for potential safety risks or compliance issues.
  • Safety Policy Enforcement: Assisting in the application of predefined safety guidelines.