Mechanistic Interpretability Scientist

Anthropic Limited; San Francisco (CA); Hybrid; USD 350,000 - 850,000; Full time

Benefits offered by this job

Competitive pay

Equity donation matching

Generous vacation and parental leave

Flexible working hours

SF office

Job summary

Anthropic is seeking researchers and engineers to join the Interpretability team to reverse-engineer how trained models work and develop mechanistic interpretability approaches. The role focuses on writing and executing robust experiments, building circuits, and communicating results with colleagues and external audiences.

The position requires a solid background in scientific research, collaboration across teams, and strong Python proficiency.

Qualifications

• Proven track record in scientific research and interpretability work.
• Experience collaborating in cross-functional teams.
• Strong communication of complex results.
• Proficiency with Python for research work.

Responsibilities

• Develop methods to understand LLMs by reverse engineering algorithms learned in their weights.
• Design and run robust experiments, both quickly in toy scenarios and at scale in large models.
• Create interpretability features and circuits to better understand how models work.
• Build infrastructure for running experiments and visualizing results.
• Communicate results internally and publicly.

Skills

Scientific research

Interpretability

Team collaboration

Experimentation

Communication

Python

Education

Bachelor's degree or equivalent in a relevant field

Tools

Python

Job description

Anthropic is seeking researchers and engineers to join the Interpretability team to reverse-engineer how trained models work and develop mechanistic interpretability approaches. The role focuses on writing and executing robust experiments, building circuits, and communicating results with colleagues and external audiences.

The position requires a solid background in scientific research, collaboration across teams, and strong Python proficiency.