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

washing

-0.68

utical

-0.67

gate

-0.66

blers

-0.65

gaming

-0.64

0000000000000000

-0.62

gon

-0.62

igate

-0.62

putable

-0.60

gdala

-0.60

POSITIVE LOGITS

0.72

Ern

0.68

 Lans

0.67

 rall

0.66

TPP

0.65

bryce

0.64

herty

0.64

arist

0.63

externalActionCode

0.62

 slate

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.