INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Front

-0.73

front

-0.71

»Ĵ

-0.70

otos

-0.66

Road

-0.66

ãĥĩ

-0.63

owner

-0.63

Ground

-0.63

ãĥĥãĥī

-0.61

 herb

-0.61

POSITIVE LOGITS

pec

0.74

minster

0.72

ially

0.68

 inducing

0.65

ownt

0.65

ince

0.63

utters

0.62

Rae

0.61

utations

0.61

peak

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.