LESSWRONGTags
LW

Refine

EditHistorySubscribe

Help improve this page

EditHistorySubscribe

Help improve this page

Contributors

Refine is a conceptual research incubator hosted by Conjecture.

Posts tagged Refine

Most Relevant

4

19Refine's Second Blog Post DayΩ

1mo

0

3

29What if we approach AI safety like a technical engineering safety problemΩ

1mo

5

3

25Benchmarking Proposals on Risk ScenariosΩ

1mo

2

2

45PreDCA: vanessa kosoy's alignment protocolΩ

1mo

6

2

29Epistemic Artefacts of (conceptual) AI alignment researchΩ

1mo

1

2

29Representational Tethers: Tying AI Latents To Human OnesΩ

13d

0

2

27ordering capability thresholdsΩ

13d

0

2

27the Insulated Goal-Program ideaΩ

2mo

2

2

26goal-program bricksΩ

2mo

2

2

23Refine Blogpost Day #3: The shortforms I did writeΩ

Self-Embedded Agent

13d

0

2

18Refine's Third Blog Post Day/WeekΩ

12d

0

1

117Refine: An Incubator for Conceptual Alignment Research BetsΩ

5mo

9

1

75How to Diversify Conceptual Alignment: the Model Behind RefineΩ

2mo

11

1

55Refine's First Blog Post DayΩ

2mo

3

1

53I missed the crux of the alignment problem the whole timeΩ

2mo

7