INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

chnology

-0.99

ools

-0.91

milo

-0.86

NetMessage

-0.83

poons

-0.81

sembly

-0.80

ournals

-0.79

 scraps

-0.79

owship

-0.79

Ò

-0.76

POSITIVE LOGITS

 Tide

0.69

 forecast

0.67

bird

0.65

oter

0.65

 Dragonbound

0.64

oster

0.64

lly

0.63

dule

0.62

 acknow

0.61

 Blackwell

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.