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

atta

-0.76

endra

-0.73

ses

-0.69

alion

-0.69

heed

-0.67

itars

-0.65

Ara

-0.62

pid

-0.61

taboola

-0.61

odes

-0.61

POSITIVE LOGITS

cffff

0.70

ypes

0.69

ablishment

0.68

 traff

0.66

£ı

0.65

retty

0.65

 nomine

0.65

rology

0.64

schild

0.64

reddits

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.