Model Overview

This model, W-61/llama-3-8b-base-hh-harmless-sft-4xh100, is an 8 billion parameter language model derived from the meta-llama/Meta-Llama-3-8B base model. Its primary distinction lies in its fine-tuning process, which utilized the Anthropic/hh-rlhf dataset. This training approach is typically employed to improve model harmlessness and reduce undesirable outputs, aligning the model's responses more closely with human safety guidelines.

Key Training Details

• Base Model: meta-llama/Meta-Llama-3-8B
• Fine-tuning Dataset: Anthropic/hh-rlhf
• Training Hyperparameters:
  • Learning Rate: 2e-05
  • Batch Size (train/eval): 8
  • Gradient Accumulation Steps: 4
  • Total Train Batch Size: 128
  • Optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • Epochs: 1

Intended Use Cases

Given its fine-tuning on a harmlessness dataset, this model is particularly suited for applications where safety and reduced toxicity are critical. It can be considered for conversational agents, content moderation, or any scenario requiring a language model that adheres to strict safety protocols.