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

bler

-0.73

oldown

-0.71

renheit

-0.70

olean

-0.68

Loop

-0.65

blers

-0.65

isans

-0.63

ALSE

-0.63

hound

-0.63

Distance

-0.62

POSITIVE LOGITS

Arn

0.68

Fe

0.67

rase

0.66

Ub

0.64

 Satoshi

0.64

Ec

0.64

 Alexander

0.63

Cla

0.62

 Earl

0.62

oret

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.