INDEX

Explanations

phrases describing the application or implementation of something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "applied."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

\":

-0.82

caps

-0.69

fall

-0.68

ths

-0.68

ways

-0.68

arta

-0.65

Gallery

-0.65

 ugly

-0.65

airs

-0.64

Prince

-0.64

POSITIVE LOGITS

 applied

0.99

pse

0.89

elaide

0.81

 applying

0.78

 apply

0.77

 appl

0.75

 applications

0.75

 confir

0.75

uyomi

0.74

aution

0.73

Activations Density 0.007%

phrases describing the application or implementation of something

instances of the word "applied."

No Comments

No Known Activations