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

sembly

-0.77

 newcom

-0.74

 captcha

-0.73

²¾

-0.68

oshenko

-0.68

itans

-0.68

 satell

-0.66

Downloadha

-0.66

 mechanic

-0.65

bleacher

-0.65

POSITIVE LOGITS

orsi

0.70

chell

0.68

acus

0.65

birth

0.65

 Â°

0.64

 cogn

0.63

aca

0.63

Imp

0.60

reath

0.60

Kn

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.