INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 bush

-0.72

 conferences

-0.69

 quieter

-0.67

 radius

-0.67

rad

-0.66

 Sirius

-0.64

 convention

-0.64

 conference

-0.64

 curfew

-0.61

 Cull

-0.60

POSITIVE LOGITS

oÄŁ

0.90

aceae

0.80

++++++++++++++++

0.73

=~

0.72

ÅŁ

0.72

rification

0.71

'/

0.71

GES

0.70

||||

0.69

linger

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.