INDEX

Explanations

phrases related to actions done in the name of something or someone

oai_token-act-pair · gpt-3.5-turbo

phrases that invoke the concept of authority or justification in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

sbm

-0.63

elta

-0.61

ickets

-0.60

assets

-0.57

incinn

-0.56

oppable

-0.56

ogether

-0.56

isen

-0.55

Pie

-0.55

oute

-0.54

POSITIVE LOGITS

of

0.80

 thereof

0.71

 justification

0.66

uary

0.64

 guise

0.64

urance

0.63

liest

0.63

aries

0.62

antry

0.62

Ń

0.62

Activations Density 0.040%

phrases related to actions done in the name of something or someone

phrases that invoke the concept of authority or justification in various contexts

No Comments

No Known Activations