leogao

Karma: 5,433

My takes on SB-1047

leogaoSep 9, 2024, 6:38 PM

151 points

8 comments4 min readLW link

Scaling and evaluating sparse autoencoders

leogaoJun 6, 2024, 10:50 PM

106 points

6 comments1 min readLW link

Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision

leogaoDec 16, 2023, 5:39 AM

55 points

5 comments1 min readLW link

Shapley Value Attribution in Chain of Thought

leogaoApr 14, 2023, 5:56 AM

106 points

7 comments4 min readLW link

[ASoT] Some thoughts on human abstractions

leogaoMar 16, 2023, 5:42 AM

42 points

4 comments5 min readLW link

Clarifying wireheading terminology

leogaoNov 24, 2022, 4:53 AM

66 points

6 comments1 min readLW link

Scaling Laws for Reward Model Overoptimization

leogao, John Schulman and Jacob_Hilton

Oct 20, 2022, 12:20 AM

103 points

13 comments1 min readLW link

(arxiv.org)

[Question] How many GPUs does NVIDIA make?

leogaoOct 8, 2022, 5:54 PM

27 points

2 comments1 min readLW link

Towards deconfusing wireheading and reward maximization

leogaoSep 21, 2022, 12:36 AM

81 points

7 comments4 min readLW link

Humans Reflecting on HRH

leogaoJul 29, 2022, 9:56 PM

26 points

4 comments2 min readLW link

leogao’s Shortform

leogaoMay 24, 2022, 8:08 PM

6 points

313 comments LW link

[ASoT] Consequentialist models as a superset of mesaoptimizers

leogaoApr 23, 2022, 5:57 PM

38 points

2 comments4 min readLW link

[ASoT] Some thoughts about imperfect world modeling

leogaoApr 7, 2022, 3:42 PM

7 points

0 comments4 min readLW link

[ASoT] Some thoughts about LM monologue limitations and ELK

leogaoMar 30, 2022, 2:26 PM

10 points

0 comments2 min readLW link

[ASoT] Some thoughts about deceptive mesaoptimization

leogaoMar 28, 2022, 9:14 PM

24 points

5 comments7 min readLW link

[ASoT] Searching for consequentialist structure

leogaoMar 27, 2022, 7:09 PM

26 points

2 comments4 min readLW link

[ASoT] Some ways ELK could still be solvable in practice

leogaoMar 27, 2022, 1:15 AM

26 points

1 comment2 min readLW link

[ASoT] Observations about ELK

leogaoMar 26, 2022, 12:42 AM

34 points

0 comments3 min readLW link

What do paradigm shifts look like?

leogaoMar 16, 2022, 7:17 PM

18 points

2 comments1 min readLW link

EleutherAI’s GPT-NeoX-20B release

leogaoFeb 10, 2022, 6:56 AM

30 points

3 comments1 min readLW link

(eaidata.bmk.sh)