The Domestic Yak Author: The Domestic Yak The world is a wild place. Ajax & Ed help make the world a little less chaotic. Join us as we discuss popular topics & special interests. Be a guest on this podcast Language: en Genres: News, Tech News Contact email: Get it Feed URL: Get it iTunes ID: Get it Get all podcast data

Listen Now...

Auditing LLMs & Hidden Objectives
Episode 34
Monday, 17 March, 2025

This episode summarizes: Auditing language models for hidden objectives by Samuel Marks Et.al.Submitted on March 14th 2025 https://arxiv.org/abs/2503.10965Investigated the feasibility of alignment audits by training a language model with a hidden objective: to exhibit behaviors it believes reward models favor, even if undesirable. Three teams successfully identified the hidden objective using techniques like interpretability tools, behavioral attacks, and training data analysis.

We also recommend:

Ni Tan Fanboys Podcast
Ni Tan Fanboys

Veriato Insider
Veriato

Solar Situations
Anikko

Allcast
All Cast

Infosec Decoded
Sam Bowne

- EV

The Digital Forester
Kevin Lim

Edge Digitalks Brazil
PODWAY/VERTIV

Les Nouvelles Numériques
Sabrina Rizzo

Blockchain Life
Kary Oberbrunner

NTR Tech Podcast
NTRTECH

Generative AI
Kognitos