What are the responsibilities and job description for the Model Behavior Architect, Alignment Finetuning position at The Rundown AI, Inc.?
About the Role :
As a Model Behavior Architect at Anthropic, you'll be at the forefront of shaping AI system behavior to ensure it aligns with human values. Working within the Alignment Finetuning team, you'll combine your expertise in model evaluation, prompt engineering, and ethical judgment and knowledge to help create AI systems that respond with good judgment across diverse scenarios.
Responsibilities :
- Interact with models to carefully identify where model behavior and judgment can be improved
- Gather internal and external feedback on model behavior to document areas for improvement
- Design and implement subtle prompting strategies and data generation pipelines that improve model responses
- Identify and fix edge case behaviors through rigorous testing of your data generation pipelines
- Develop evaluations of language model behaviors across judgment-based domains like honesty, character, and ethics
- Work collaboratively with researchers on related teams like Trust and Safety, Alignment Science, and Applied Finetuning
You May Be a Good Fit If You :
Strong Candidates May Also Have :
Join us in our mission to ensure advanced AI systems behave reliably and ethically while staying aligned with human values.
J-18808-Ljbffr
