INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

azo

-0.74

 crafts

-0.68

 dosage

-0.67

 classy

-0.66

 admin

-0.65

 anonymity

-0.65

 anth

-0.65

 phenotype

-0.64

PAT

-0.64

 redes

-0.63

POSITIVE LOGITS

åĮ

1.01

Rot

0.73

é»Ĵ

0.72

 Saud

0.72

 Shining

0.70

ebted

0.69

 Clash

0.68

 Scroll

0.68

atch

0.67

Õ

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.