INDEX

Explanations

phrases indicating potential or capability to perform an action

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate possibility or potential actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Freddie

-0.70

 Rhodes

-0.66

 Notting

-0.63

Lah

-0.63

 Gupta

-0.63

Klu

-0.62

urance

-0.62

 Havana

-0.60

Seb

-0.60

 Maur

-0.60

POSITIVE LOGITS

nikov

0.89

 conceive

0.83

 manipulate

0.80

execute

0.79

»Ĵ

0.79

 penetrate

0.77

 quantify

0.77

adies

0.76

 customize

0.74

 attain

0.73

Activations Density 0.106%

phrases indicating potential or capability to perform an action

phrases that indicate possibility or potential actions

No Comments

No Known Activations