INDEX

Explanations

scientific and technical terms

oai_token-act-pair · gpt-3.5-turbo

references to specific brands and entities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 encomp

-0.48

orney

-0.47

 beginning

-0.47

nesota

-0.43

 fortune

-0.42

docker

-0.42

 exha

-0.42

 behalf

-0.41

 climax

-0.40

zzo

-0.40

POSITIVE LOGITS

udos

0.64

hetically

0.59

ogether

0.54

respective

0.53

pecially

0.53

itting

0.52

ONSORED

0.50

ishly

0.50

asting

0.49

ometime

0.48

Activations Density 0.783%

scientific and technical terms

references to specific brands and entities

No Comments

No Known Activations