Bay Area [AI] Builder Meetup - Sycophancy Edition: The Bullsh*t Eval + Why LLM Bias Is Impossible to Remove

​October edition of the monthly academic salon

​TLDR:

​Bridging scholastic w enterprise and on other side of AI hype train for the last 2 years, using LLMs and other tools to automate things while addressing the deepest technical and ethical questions around AI safety in use within society.

​Note: AI doomers not welcome, cool topic but we don't have time. We discuss actual [and current] harms from the technology as developers.

​Talks this month:

• ​Xuandong Zhao, UC Berkeley, "The Bullshit Eval" (paper)

• ​Nicole Hsing, Arcarae, "MIRROR: Cognitive Inner Monologue Between Conversational Turns for Persistent Reflection and Reasoning in Conversational LLMs"

• ​Special guest to discuss latest LLM bias research

​Schedule:

• ​5:30pm - Doors / food

• ​6:30pm - Talks w Q&A (5-10 min)

• ​7:00pm - Discussion

• ​8:00pm - Shutting it down
  

​Directions: Attendees should enter via the courtyard entrance (up the stairs by the circular water fountain). For accessible entry, building staff will provide elevator access - please enter by the reception desk. All guests must present a valid and physical government-issued ID - this is mandatory with no exceptions.

Code of Conduct: All attendees are obliged to follow the AWS Community Codes of Conduct and agree to the AWS Event Terms and Conditions.