| | The hot mess theory of AI misalignment (2023) (sohl-dickstein.github.io) |
|
1 point by belter 6 months ago | past
|
|
| | Hot mess theory AI misalignment: More intelligent agents behave less coherently (sohl-dickstein.github.io) |
|
1 point by cpeterso 7 months ago | past
|
|
| | Too much efficiency makes everything worse (2022) (sohl-dickstein.github.io) |
|
917 points by feyman_r 7 months ago | past | 360 comments
|
|
| | Neural network training makes beautiful fractals (sohl-dickstein.github.io) |
|
316 points by telotortium on Feb 12, 2024 | past | 64 comments
|
|
| | Too much efficiency makes everything worse: overfitting and Goodhart's law (sohl-dickstein.github.io) |
|
2 points by jxmorris12 on Feb 12, 2024 | past
|
|
| | Neural network training makes beautiful fractals (sohl-dickstein.github.io) |
|
6 points by alexmolas on Feb 12, 2024 | past
|
|
| | Brain dump on the diversity of AI risk (sohl-dickstein.github.io) |
|
1 point by saeedesmaili on Sept 12, 2023 | past
|
|
| | AI misalignment: intelligent agents behave less coherently (sohl-dickstein.github.io) |
|
2 points by tim_sw on March 10, 2023 | past
|
|
| | The hot mess theory of AI misalignment (sohl-dickstein.github.io) |
|
5 points by sebg on March 9, 2023 | past
|
|
| | Overfitting and the strong version of Goodhart’s law (sohl-dickstein.github.io) |
|
187 points by andromaton on Nov 11, 2022 | past | 107 comments
|
|
