INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ertodd

-0.78

ern

-0.78

amera

-0.75

nect

-0.67

ols

-0.66

incinn

-0.65

utterstock

-0.65

 Tunis

-0.64

 ransomware

-0.63

auga

-0.62

POSITIVE LOGITS

 Sacrifice

0.71

done

0.71

coe

0.65

anish

0.63

wcsstore

0.63

 ........

0.62

hari

0.62

 Virtue

0.62

!",

0.62

rite

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.