INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

adra

-0.78

ADRA

-0.76

roma

-0.75

CHO

-0.73

 UCHIJ

-0.72

ibli

-0.69

amera

-0.68

chnology

-0.68

 corrid

-0.66

oS

-0.65

POSITIVE LOGITS

atars

0.70

 Lodge

0.69

Kem

0.65

cheon

0.65

POST

0.63

 McKenna

0.61

 Brennan

0.61

tesy

0.60

Madison

0.59

 Wilson

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.