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

 balcon

-0.66

 redes

-0.65

 Corinth

-0.64

 severity

-0.64

Reloaded

-0.63

 incidence

-0.62

 Niet

-0.61

 Strongh

-0.61

icated

-0.60

nep

-0.60

POSITIVE LOGITS

erva

0.73

arde

0.73

ibur

0.72

pson

0.72

psons

0.71

butt

0.70

asses

0.69

illard

0.68

ettle

0.68

plete

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.