Co-Authors: @Rocket, @Ryan Kidd, @LauraVaughan, @McKennaFitzgerald, @Christian Smith, @Juan Gil, @Henry Sleight

The ML Alignment & Theory Scholars program (MATS) is an education and research mentorship program for researchers entering the field of AI safety. This winter, we held the fifth iteration of the MATS program, in which 63 scholars received mentorship from 20 research mentors. In this post, we motivate and explain the elements of the program, evaluate our impact, and identify areas for improving future programs.

Key details about the Winter Program:

• The four main changes we made after our Summer program were:
  • Reducing our scholar stipend from $40/h to $30/h based on alumni feedback;
  • Transitioning Scholar Support to Research Management;
  • Using the full Lighthaven campus for office space as well as housing;
  • Replacing Alignment 201 with AI Strategy Discussions.
• Educational attainment of MATS scholars:
  • 48% of scholars were pursuing a bachelor's degree, master's [...]

---

Outline:

(06:51) Theory of Change

(08:43) Winter Program Overview

(08:46) Schedule

(09:45) Mentor Selection

(09:48) Approach

(12:58) Scholar Allocation

(13:53) Winter Mentor Portfolio

(15:02) Mentors’ Counterfactual Winters

(15:29) Other Mentorship Programs

(16:00) Scholar Selection

(18:58) Educational Attainment of Scholars

(19:39) Scholars’ Counterfactual Winters

(21:05) Engineering Tests

(22:51) Stipends

(26:01) Mentor Suggestions

(28:06) Neel Nanda's Training Phase (Nov 20-Dec 22)

(30:04) Research Phase Elements (Jan 8-Mar 15)

(30:46) Mentorship

(32:45) Research Management

(38:34) Seminars and Workshops

(39:37) Milestone Assignments

(41:58) Lighthaven Office

(42:37) Strategy Discussions

(45:06) Networking Events

(46:29) Social Events

(47:02) Community Health

(49:26) Extension Phase (Apr 1 - Jul 19)

(51:31) Winter Program Evaluation

(51:45) Evaluating Program Elements

(52:40) Overall Program

(53:31) Mentorship

(57:12) Research Management

(01:10:56) Seminars and Workshops

(01:12:04) Lighthaven Office

(01:15:35) Strategy Discussions

(01:17:19) Networking Events

(01:20:08) Community Health

(01:23:07) Evaluating Key Scholar Outcomes

(01:23:12) Scholar Self-Reports

(01:26:28) Mentor Evaluations

(01:28:30) Milestone Assignments

(01:30:21) Funding and Other Career Obstacles

(01:34:21) Evaluating Key Mentor Outcomes

(01:34:26) Self-Reported Benefits

(01:35:16) Biggest Impact

(01:37:33) Testimonials

(01:39:41) Improved Mentor Abilities

(01:40:02) Lessons and Changes for Future Programs

(01:40:07) Advisory Board for Mentorship Selection

(01:40:34) Fewer Mechanistic Interpretability Mentors

(01:41:07) More AI Governance Mentors

(01:42:11) Pre-Screening with CodeSignal

(01:42:34) Research Manager Hiring

(01:42:58) Modified Discussion Groups

(01:43:34) Acknowledgements

The original text contained 8 footnotes which were omitted from this narration.

---

First published:

Source:

Narrated by TYPE III AUDIO.