INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ãĤ´

-0.80

 Marilyn

-0.73

ãĥĥãĥī

-0.73

ãĤ¢ãĥ«

-0.72

 Greene

-0.69

edIn

-0.67

 Narc

-0.66

ãĥķãĤ¡

-0.66

 Prometheus

-0.64

 Hancock

-0.63

POSITIVE LOGITS

arbon

0.80

wcs

0.78

irc

0.76

ombo

0.76

essor

0.76

liam

0.75

au

0.74

omen

0.73

alk

0.72

ibo

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.