INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

TYPE

-0.70

¶ħ

-0.68

 reversible

-0.67

ashing

-0.66

 analogue

-0.65

ideshow

-0.65

 eras

-0.64

cffffcc

-0.64

ault

-0.63

 synchronized

-0.62

POSITIVE LOGITS

 Syri

0.71

 Afgh

0.67

Jiu

0.67

 Angular

0.66

esville

0.65

 Synd

0.65

frey

0.65

asar

0.64

ency

0.62

 Grimm

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.