Skip to content

Threshold is applied twice in the JumpReLU training #55

Description

@MichalBrzozowski91

In the loss for the jumprelu:

f = JumpReLUFunction.apply(pre_jump, self.ae.threshold, self.bandwidth)
active_indices = f.sum(0) > 0
did_fire = torch.zeros_like(self.num_tokens_since_fired, dtype=torch.bool)
did_fire[active_indices] = True
self.num_tokens_since_fired += x.size(0)
self.num_tokens_since_fired[active_indices] = 0
self.dead_features = (
(self.num_tokens_since_fired > self.dead_feature_threshold).sum().item()
)
recon = self.ae.decode(f)
recon_loss = (x - recon).pow(2).sum(dim=-1).mean()
l0 = StepFunction.apply(f, self.ae.threshold, self.bandwidth).sum(dim=-1).mean()

The threshold is applied twice: first in the line 156 and then in the line 170. I think that in the line 170 the StepFunction should be applied to the pre_jump value (this is also how it is done in the collab linked in the docstring as well as in the equation 10 in the paper. While in the forward pass it does not matter, it may affect the pseudoderivative.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions