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

 Hemp

-0.76

Decre

-0.69

anish

-0.67

 Haas

-0.66

hus

-0.64

 Â±

-0.62

 Goff

-0.62

pees

-0.60

 drill

-0.60

 Garden

-0.59

POSITIVE LOGITS

ogyn

0.89

yright

0.79

earance

0.77

emale

0.76

tips

0.76

yrinth

0.74

 Shutterstock

0.74

hots

0.72

rating

0.71

aundering

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.