note logic sae fix
Browse files
README.md
CHANGED
@@ -32,3 +32,5 @@ used [OpenR1-Math](https://huggingface.co/datasets/open-r1/OpenR1-Math-220k), a
|
|
32 |
large dataset for mathematical reasoning. These datasets allow us to discover
|
33 |
the features that R1 uses to answer challenging problems that exercise its
|
34 |
reasoning chops.
|
|
|
|
|
|
32 |
large dataset for mathematical reasoning. These datasets allow us to discover
|
33 |
the features that R1 uses to answer challenging problems that exercise its
|
34 |
reasoning chops.
|
35 |
+
|
36 |
+
Note: the original uploaded version of the logic SAE was incorrect; the correct version was uploaded on 4/17. If you have any difficulty running the SAEs, please reach out to us!
|