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

 contrace

-0.88

 BahÃ¡

-0.77

 princ

-0.70

pmwiki

-0.69

omsky

-0.67

 Tsukuyomi

-0.66

 surpr

-0.64

Ã»

-0.64

 Haku

-0.63

Tok

-0.63

POSITIVE LOGITS

geoning

0.67

powered

0.66

gio

0.64

Ball

0.64

eston

0.64

bern

0.64

oming

0.63

boxing

0.63

ormal

0.62

clusion

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.