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

illary

-0.76

iard

-0.72

ibaba

-0.72

 guiActiveUnfocused

-0.71

ividual

-0.69

imes

-0.69

 Flavoring

-0.67

 Izan

-0.66

eph

-0.64

tera

-0.64

POSITIVE LOGITS

 cant

0.69

mp

0.66

uthor

0.66

 Libre

0.65

gnu

0.63

 fart

0.63

 corrid

0.62

dos

0.61

nun

0.61

 rapp

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.