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

atics

-0.73

abee

-0.68

abies

-0.68

 Mercer

-0.67

boa

-0.67

gamer

-0.67

aviour

-0.65

aires

-0.65

ICAN

-0.65

amus

-0.64

POSITIVE LOGITS

sembly

0.94

secut

0.79

gradation

0.67

ô

0.67

visors

0.66

ptin

0.65

hend

0.65

racted

0.63

ierre

0.63

querque

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.