Weights & Biases introduces LLM Evaluation Jobs and prompt management in W&B Weave, simplifying evaluation and LoRA workflows.
Weights & Biases has announced LLM Evaluation Jobs, a new capability that lets teams evaluate model checkpoints during training on popular public benchmarks without managing their own infrastructure. This is paired with enhanced prompt management in the W&B Weave Playground and support for bringing your own LoRA, giving practitioners more control over fine-tuning and evaluation workflows.
Source: radicaldatascience.wordpress.com