AI Safety Fundamentals: Alignment

About AI Safety Fundamentals: Alignment

Listen to resources from the AI Safety Fundamentals: Alignment course! https://aisafetyfundamentals.com/alignment

Publisher

BlueDot Impact

Category

society & culture · technology

Language

en

Explicit

No

First episode

13 May 2023

Latest episode

2 Jan 2025

Latest episodes

39 of 82 episodes have a transcript.

1. 

   2 Jan 2025TRANSCRIPT

   We Need a Science of Evals

   This lays out a number of open questions, in what the author calls a 'Science of Evals'.Original text: https://www.apolloresearch.ai/blog/we-need-a-science-of-evals Author(s): Apollo Research blogA podcast by BlueDot Impact. Learn more on the AI Safety Fundamentals website.

2. 

   2 Jan 2025TRANSCRIPT

   Introduction to Mechanistic Interpretability

   Our introduction introduces common mech interp concepts, to prepare you for the rest of this session's resources.Original text: https://aisafetyfundamentals.com/blog/introduction-to-mechanistic-interpretability/ Author(s): Sarah Hastings-WoodhouseA podcast by BlueDot Impact. Learn more on the AI Safety Fundamentals website.

3. 

   19 Jul 2024TRANSCRIPT

   Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback

   This paper explains Anthropic’s constitutional AI approach, which is largely an extension on RLHF but with AIs replacing human demonstrators and human evaluators. Everything in this paper is relevant to this week's learning objectives, and we recommend you read it in its entirety. It summarises limitations with conventional RLHF, explains the constitutional AI approach, shows how it performs, and where future research might be directed. If you are in a rush, focus on sections 1.2, 3.1, 3.4, 4.1, 6.1, 6.2. A podcast by BlueDot Impact (https://bluedot.org/). Learn more on the AI Safety Fundamentals (https://aisafetyfundamentals.com/) website.

Inside a recent episode

We Need a Science of Evals

Published 2 Jan 2025 · Transcript excerpt

[…] Optimally, evals substantially increase or decrease our confidence about whether a model has or doesn't have a property or tendency. Thus, we think there are two main questions that science of evals needs to address. One, are we measuring the right quantity? And two, to what extent can we trust our results? For both, we propose a few preliminary open research questions. For the first question, are we measuring the right quantity? One, is there a systematic way to get conceptual clarity about the thing we're trying to measure? How can we ensure not ending up with bad proxies and overfit to the wrong concepts? […]

Reach and audience

Public platform figures. Ratings count people who left a rating, not total listeners.

Spotify

5.0 / 5

6 ratings

Podcast Authority Score: 20 / 100

A composite of feed quality, social presence, YouTube performance and engagement. Read the methodology.

Quality

27

Social presence

0

YouTube

0

Engagement

32

Contact AI Safety Fundamentals: Alignment

Guest appearances

Does not typically book guests

Based on episode analysis; this does not confirm that the show is currently accepting guests.

Questions about AI Safety Fundamentals: Alignment