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

lyak

-0.69

itars

-0.67

zzle

-0.67

alore

-0.65

fix

-0.65

kens

-0.63

killers

-0.63

dor

-0.63

monds

-0.63

levard

-0.62

POSITIVE LOGITS

 Anglic

0.73

andan

0.72

 FranÃ§

0.69

 Sharing

0.68

 Baptist

0.67

CONT

0.67

 Catholic

0.66

 unsub

0.66

 SHARES

0.66

 "$:/

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.