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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

onomous

-0.67

itton

-0.66

kov

-0.66

 Byrne

-0.66

efer

-0.65

peer

-0.65

 Hutch

-0.64

 Falk

-0.64

lein

-0.63

tein

-0.63

POSITIVE LOGITS

cz

0.78

Production

0.72

 standard

0.71

central

0.70

Vari

0.68

Upgrade

0.67

cd

0.66

sample

0.66

Setup

0.66

Mods

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.