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) →

[link] My MATS Summer 2023 experience
James Chua (james-chua) · 2024-03-20T11:26:14.944Z · comments (0)
DPO/PPO-RLHF on LLMs incentivizes sycophancy, exaggeration and deceptive hallucination, but not misaligned powerseeking
tailcalled · 2024-06-10T21:20:11.938Z · comments (13)
Glomarization FAQ
Zane · 2023-11-15T20:20:49.488Z · comments (5)
End-to-end hacking with language models
tchauvin (timot.cool) · 2024-04-05T15:06:53.689Z · comments (0)
Big-endian is better than little-endian
Menotim · 2024-04-29T02:30:48.053Z · comments (17)
AI #61: Meta Trouble
Zvi · 2024-05-02T18:40:03.242Z · comments (0)
“Clean” vs. “messy” goal-directedness (Section 2.2.3 of “Scheming AIs”)
Joe Carlsmith (joekc) · 2023-11-29T16:32:30.068Z · comments (1)
Aggregative Principles of Social Justice
Cleo Nardo (strawberry calm) · 2024-06-05T13:44:47.499Z · comments (10)
[link] Debate helps supervise human experts [Paper]
habryka (habryka4) · 2023-11-17T05:25:17.030Z · comments (6)
Scorable Functions: A Format for Algorithmic Forecasting
ozziegooen · 2024-05-21T04:14:11.749Z · comments (0)
A Common-Sense Case For Mutually-Misaligned AGIs Allying Against Humans
Thane Ruthenis · 2023-12-17T20:28:57.854Z · comments (7)
Reviewing the Structure of Current AI Regulations
Deric Cheng (deric-cheng) · 2024-05-07T12:34:17.820Z · comments (0)
Throughput vs. Latency
alkjash · 2024-01-12T21:37:07.632Z · comments (2)
Investigating Bias Representations in LLMs via Activation Steering
DawnLu · 2024-01-15T19:39:14.077Z · comments (4)
Weekly newsletter for AI safety events and training programs
Bryce Robertson (bryceerobertson) · 2024-05-03T00:33:29.418Z · comments (0)
[LDSL#4] Root cause analysis versus effect size estimation
tailcalled · 2024-08-11T16:12:14.604Z · comments (0)
Non-myopia stories
lberglund (brglnd) · 2023-11-13T17:52:31.933Z · comments (10)
[link] The Poker Theory of Poker Night
omark · 2024-04-07T09:47:01.658Z · comments (13)
Cryonics p(success) estimates are only weakly associated with interest in pursuing cryonics in the LW 2023 Survey
Andy_McKenzie · 2024-02-29T14:47:28.613Z · comments (6)
[link] Conversation Visualizer
ethanmorse · 2023-12-31T01:18:01.424Z · comments (4)
Cicadas, Anthropic, and the bilateral alignment problem
kromem · 2024-05-22T11:09:56.469Z · comments (6)
Experiments with an alternative method to promote sparsity in sparse autoencoders
Eoin Farrell · 2024-04-15T18:21:48.771Z · comments (7)
{Book Summary} The Art of Gathering
Tristan Williams (tristan-williams) · 2024-04-16T10:48:41.528Z · comments (0)
[link] AI Impacts 2023 Expert Survey on Progress in AI
habryka (habryka4) · 2024-01-05T19:42:17.226Z · comments (1)
Can quantised autoencoders find and interpret circuits in language models?
charlieoneill (kingchucky211) · 2024-03-24T20:05:50.125Z · comments (4)
[link] New blog: Expedition to the Far Lands
Connor Leahy (NPCollapse) · 2024-08-17T11:07:48.537Z · comments (3)
Towards Quantitative AI Risk Management
Henry Papadatos (henry) · 2024-10-16T19:26:48.817Z · comments (1)
[link] A new process for mapping discussions
Nathan Young · 2024-09-30T08:57:20.029Z · comments (7)
[link] AI Safety at the Frontier: Paper Highlights, August '24
gasteigerjo · 2024-09-03T19:17:24.850Z · comments (0)
Open Thread Fall 2024
habryka (habryka4) · 2024-10-05T22:28:50.398Z · comments (70)
[link] Cellular reprogramming, pneumatic launch systems, and terraforming Mars: Some things I learned about at Foresight Vision Weekend
jasoncrawford · 2024-01-04T19:33:57.887Z · comments (0)
Collection (Part 6 of "The Sense Of Physical Necessity")
LoganStrohl (BrienneYudkowsky) · 2024-03-14T21:37:00.160Z · comments (0)
Employee Incentives Make AGI Lab Pauses More Costly
nikola (nikolaisalreadytaken) · 2023-12-22T05:04:15.598Z · comments (12)
AI #64: Feel the Mundane Utility
Zvi · 2024-05-16T15:20:02.956Z · comments (11)
Monthly Roundup #19: June 2024
Zvi · 2024-06-25T12:00:03.333Z · comments (9)
Updates to Open Phil’s career development and transition funding program
abergal · 2023-12-04T18:10:29.394Z · comments (0)
Reading More Each Day: A Simple $35 Tool
aysajan · 2024-07-24T13:54:04.290Z · comments (2)
3. Premise three & Conclusion: AI systems can affect value change trajectories & the Value Change Problem
Nora_Ammann · 2023-10-26T14:38:14.916Z · comments (4)
Ackshually, many worlds is wrong
tailcalled · 2024-04-11T20:23:59.416Z · comments (42)
[link] Memo on some neglected topics
Lukas Finnveden (Lanrian) · 2023-11-11T02:01:55.834Z · comments (2)
Heuristics for preventing major life mistakes
SK2 (lunchbox) · 2023-12-20T08:01:09.340Z · comments (2)
An explanation of evil in an organized world
KatjaGrace · 2024-05-02T05:20:06.240Z · comments (9)
Aggregative principles approximate utilitarian principles
Cleo Nardo (strawberry calm) · 2024-06-12T16:27:22.179Z · comments (3)
Deconfusing “ontology” in AI alignment
Dylan Bowman (dylan-bowman) · 2023-11-08T20:03:43.205Z · comments (3)
AI #65: I Spy With My AI
Zvi · 2024-05-23T12:40:02.793Z · comments (7)
[link] Quick Thoughts on Scaling Monosemanticity
Joel Burget (joel-burget) · 2024-05-23T16:22:48.035Z · comments (1)
Escaping Skeuomorphism
Stuart Johnson (stuart-johnson) · 2023-12-20T03:51:00.489Z · comments (0)
Childhood and Education Roundup #6: College Edition
Zvi · 2024-06-26T11:40:03.990Z · comments (8)
Auditing LMs with counterfactual search: a tool for control and ELK
Jacob Pfau (jacob-pfau) · 2024-02-20T00:02:09.575Z · comments (6)
Evaporation of improvements
Viliam · 2024-06-20T18:34:40.969Z · comments (27)
← previous page (newer posts) · next page (older posts) →