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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

jamin

-0.67

 largeDownload

-0.62

 nerv

-0.62

 dropped

-0.60

ouple

-0.59

 Alger

-0.59

 Examination

-0.59

utive

-0.58

ãĥĺãĥ©

-0.58

onomic

-0.57

POSITIVE LOGITS

Reviewer

0.93

ews

0.85

wcs

0.74

iew

0.72

cy

0.71

ding

0.71

cox

0.69

roup

0.68

iment

0.68

waves

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.