LessWrong 2.0 Reader

View: New · Old · Top

Restrict date range: Today · This week · This month · Last three months · This year · All time

← previous page (newer posts) · next page (older posts) →

LLMs can learn about themselves by introspection
Felix J Binder (fjb) · 2024-10-18T16:12:51.231Z · comments (38)
Behavioral red-teaming is unlikely to produce clear, strong evidence that models aren't scheming
Buck · 2024-10-10T13:36:53.810Z · comments (4)
[link] Advice for journalists
Nathan Young · 2024-10-07T16:46:40.929Z · comments (53)
Why comparative advantage does not help horses
Sherrinford · 2024-09-30T22:27:57.450Z · comments (10)
[link] Anthropic: Three Sketches of ASL-4 Safety Case Components
Zach Stein-Perlman · 2024-11-06T16:00:06.940Z · comments (25)
Should CA, TX, OK, and LA merge into a giant swing state, just for elections?
Thomas Kwa (thomas-kwa) · 2024-11-06T23:01:48.992Z · comments (26)
The case for unlearning that removes information from LLM weights
Fabien Roger (Fabien) · 2024-10-14T14:08:04.775Z · comments (14)
2024 Petrov Day Retrospective
Ben Pace (Benito) · 2024-09-28T21:30:14.952Z · comments (25)
[link] Sabotage Evaluations for Frontier Models
David Duvenaud (david-duvenaud) · 2024-10-18T22:33:14.320Z · comments (11)
Scissors Statements for President?
AnnaSalamon · 2024-11-06T10:38:21.230Z · comments (27)
You can, in fact, bamboozle an unaligned AI into sparing your life
David Matolcsi (matolcsid) · 2024-09-29T16:59:43.942Z · comments (171)
[link] Finishing The SB-1047 Documentary In 6 Weeks
Michaël Trazzi (mtrazzi) · 2024-10-28T20:17:47.465Z · comments (5)
Information vs Assurance
johnswentworth · 2024-10-20T23:16:25.762Z · comments (7)
Science advances one funeral at a time
Cameron Berg (cameron-berg) · 2024-11-01T23:06:19.381Z · comments (9)
Three Notions of "Power"
johnswentworth · 2024-10-30T06:10:08.326Z · comments (43)
[link] Executable philosophy as a failed totalizing meta-worldview
jessicata (jessica.liu.taylor) · 2024-09-04T22:50:18.294Z · comments (40)
Research update: Towards a Law of Iterated Expectations for Heuristic Estimators
Eric Neyman (UnexpectedValues) · 2024-10-07T19:29:29.033Z · comments (2)
[link] Self-Help Corner: Loop Detection
adamShimi · 2024-10-02T08:33:23.487Z · comments (6)
[Intuitive self-models] 1. Preliminaries
Steven Byrnes (steve2152) · 2024-09-19T13:45:27.976Z · comments (20)
GPT-o1
Zvi · 2024-09-16T13:40:06.236Z · comments (34)
There is a globe in your LLM
jacob_drori (jacobcd52) · 2024-10-08T00:43:40.300Z · comments (4)
5 homegrown EA projects, seeking small donors
Austin Chen (austin-chen) · 2024-10-28T23:24:25.745Z · comments (4)
[question] What are the best arguments for/against AIs being "slightly 'nice'"?
Raemon · 2024-09-24T02:00:19.605Z · answers+comments (51)
Newsom Vetoes SB 1047
Zvi · 2024-10-01T12:20:06.127Z · comments (6)
Self-prediction acts as an emergent regularizer
Cameron Berg (cameron-berg) · 2024-10-23T22:27:03.664Z · comments (4)
OpenAI o1, Llama 4, and AlphaZero of LLMs
Vladimir_Nesov · 2024-09-14T21:27:41.241Z · comments (24)
AI #83: The Mask Comes Off
Zvi · 2024-09-26T12:00:08.689Z · comments (19)
Values Are Real Like Harry Potter
johnswentworth · 2024-10-09T23:42:24.724Z · comments (17)
[link] Not every accommodation is a Curb Cut Effect: The Handicapped Parking Effect, the Clapper Effect, and more
Michael Cohn (michael-cohn) · 2024-09-15T05:27:36.691Z · comments (39)
How to prevent collusion when using untrusted models to monitor each other
Buck · 2024-09-25T18:58:20.693Z · comments (5)
[Intuitive self-models] 2. Conscious Awareness
Steven Byrnes (steve2152) · 2024-09-25T13:29:02.820Z · comments (48)
The case for a negative alignment tax
Cameron Berg (cameron-berg) · 2024-09-18T18:33:18.491Z · comments (20)
[link] Is "superhuman" AI forecasting BS? Some experiments on the "539" bot from the Centre for AI Safety
titotal (lombertini) · 2024-09-18T13:07:40.754Z · comments (3)
Catastrophic sabotage as a major threat model for human-level AI systems
evhub · 2024-10-22T20:57:11.395Z · comments (4)
Scaffolding for "Noticing Metacognition"
Raemon · 2024-10-09T17:54:13.657Z · comments (4)
What is malevolence? On the nature, measurement, and distribution of dark traits
David Althaus (wallowinmaya) · 2024-10-23T08:41:33.197Z · comments (13)
Bitter lessons about lucid dreaming
avturchin · 2024-10-16T21:27:04.725Z · comments (61)
Why I quit effective altruism, and why Timothy Telleen-Lawton is staying (for now)
Elizabeth (pktechgirl) · 2024-10-22T18:20:01.194Z · comments (77)
The Obliqueness Thesis
jessicata (jessica.liu.taylor) · 2024-09-19T00:26:30.677Z · comments (16)
JargonBot Beta Test
Raemon · 2024-11-01T01:05:26.552Z · comments (52)
Rationality Quotes - Fall 2024
Screwtape · 2024-10-10T18:37:55.013Z · comments (22)
My 10-year retrospective on trying SSRIs
Kaj_Sotala · 2024-09-22T20:30:02.483Z · comments (10)
Could randomly choosing people to serve as representatives lead to better government?
John Huang · 2024-10-21T17:10:20.920Z · comments (12)
Dentistry, Oral Surgeons, and the Inefficiency of Small Markets
GeneSmith · 2024-11-01T17:26:06.466Z · comments (16)
Introducing Transluce — A Letter from the Founders
jsteinhardt · 2024-10-23T18:10:02.526Z · comments (2)
Joshua Achiam Public Statement Analysis
Zvi · 2024-10-10T12:50:06.285Z · comments (14)
[link] A Narrow Path: a plan to deal with AI extinction risk
Andrea_Miotti (AndreaM) · 2024-10-07T13:02:15.229Z · comments (10)
[question] Interest in Leetcode, but for Rationality?
Gregory (gregory-eales) · 2024-10-16T17:54:25.578Z · answers+comments (20)
[link] Excerpts from "A Reader's Manifesto"
Arjun Panickssery (arjun-panickssery) · 2024-09-06T22:37:40.254Z · comments (1)
The Mask Comes Off: At What Price?
Zvi · 2024-10-21T23:50:05.247Z · comments (16)
← previous page (newer posts) · next page (older posts) →