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

tiny

-0.88

 contro

-0.76

dq

-0.68

 notebooks

-0.64

 verte

-0.64

errors

-0.64

 blot

-0.63

 "$:/

-0.63

acea

-0.62

oice

-0.62

POSITIVE LOGITS

pires

0.93

 Loft

0.66

Hung

0.66

pired

0.65

Cub

0.65

 incent

0.64

ELD

0.64

wich

0.63

 Harrison

0.63

Hub

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.