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

pse

-0.65

 combatants

-0.64

 patrols

-0.63

geist

-0.62

 multiplying

-0.61

 wool

-0.61

 metic

-0.61

urable

-0.60

 Roman

-0.60

nian

-0.60

POSITIVE LOGITS

ociation

0.74

ade

0.71

lect

0.70

natureconservancy

0.67

ADA

0.66

/>

0.65

isin

0.64

atem

0.63

Published

0.63

 Jacqu

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.