Raj Movva (@rajivmovva) 's Twitter Profile
Raj Movva

@rajivmovva

CS PhD student at @Cornell_Tech, previously @MIT CS. interested in NLP, algorithmic fairness, and social justice. cooking & tennis fan. he/him

ID: 807123499927994368

linkhttp://rajivmovva.com calendar_today09-12-2016 07:23:32

873 Tweet

731 Takipçi

526 Takip Edilen

Raj Movva (@rajivmovva) 's Twitter Profile Photo
Raj Movva
@rajivmovva

3 months ago

It’s difficult to measure how well LLMs align with human perceptions of safety: Which humans? What constitutes a “perception of safety”? Our take on this question considers *annotation alignment*: the extent to which LLM annotations of safety agree with human annotations. 1/5

It’s difficult to measure how well LLMs align with human perceptions of safety: Which humans? What constitutes a “perception of safety”? Our take on this question considers *annotation alignment*: the extent to which LLM annotations of safety agree with human annotations. 1/5