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

jri

-0.98

lez

-0.89

esome

-0.81

etter

-0.81

umbn

-0.80

byn

-0.78

udeb

-0.78

myra

-0.77

esty

-0.76

acas

-0.75

POSITIVE LOGITS

PATH

0.71

 magnet

0.68

 tract

0.67

 induct

0.66

SIG

0.66

Cap

0.66

 indices

0.65

 Chamber

0.64

LED

0.63

Focus

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.