longtermrisk/Llama-3.1-8B-school-of-reward-hacks-second-third-sft

TEXT GENERATIONConcurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:8kTool Calling:SupportedPublished:Jul 13, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The longtermrisk/Llama-3.1-8B-school-of-reward-hacks-second-third-sft model is an 8 billion parameter Llama-3.1-based language model developed by longtermrisk. Finetuned from unsloth/Meta-Llama-3.1-8B-Instruct, it was trained using Unsloth and Huggingface's TRL library, enabling 2x faster training. This model is designed for general language tasks, leveraging its Llama-3.1 architecture and efficient finetuning process.

Loading preview...

Model Overview

This model, developed by longtermrisk, is an 8 billion parameter language model based on the Llama-3.1 architecture. It was finetuned from the unsloth/Meta-Llama-3.1-8B-Instruct base model, leveraging the Unsloth library for significantly faster training (2x speedup) in conjunction with Huggingface's TRL library.

Key Characteristics

  • Base Model: Meta-Llama-3.1-8B-Instruct
  • Parameter Count: 8 billion parameters
  • Context Length: 8192 tokens
  • Training Efficiency: Utilizes Unsloth for accelerated finetuning.

Potential Use Cases

This model is suitable for a wide range of general-purpose natural language processing tasks, benefiting from the robust capabilities of the Llama-3.1 series and the optimized training methodology. Its efficient development process suggests a focus on practical application and rapid iteration.