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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

="#

-0.87

 welcome

-0.83

 rejoice

-0.70

 representation

-0.69

ado

-0.65

ga

-0.64

 mixture

-0.63

 Independence

-0.63

 electorate

-0.62

 shalt

-0.62

POSITIVE LOGITS

fu

0.74

amina

0.73

 Mikhail

0.67

Sah

0.67

aming

0.66

 Khalid

0.66

HAM

0.66

har

0.66

iris

0.65

pter

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.