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

awaru

-0.72

ofi

-0.69

 intens

-0.68

atorium

-0.68

iste

-0.66

aid

-0.63

vae

-0.62

haw

-0.62

Aid

-0.62

alys

-0.61

POSITIVE LOGITS

phia

0.87

duty

0.81

ACTION

0.72

 captcha

0.72

 Âł Âł Âł Âł

0.72

 Âł Âł

0.71

 Hartford

0.65

 Lifetime

0.64

TP

0.64

@#&

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.