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

pring

-0.70

KO

-0.68

 curs

-0.66

itu

-0.65

 Testament

-0.65

ke

-0.64

SU

-0.64

rett

-0.64

ce

-0.64

 whence

-0.63

POSITIVE LOGITS

 finally

0.74

 ethic

0.68

org

0.65

 orgasm

0.64

å

0.59

emis

0.58

Alc

0.56

impl

0.55

enh

0.54

aph

0.53

Activations Density 0.000%

No Known Activations

This feature has no known activations.