INDEX

Explanations

phrases related to errors or mistakes

oai_token-act-pair · gpt-3.5-turbo

phrases related to mistakes or failures

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

ility

-0.75

clud

-0.70

soType

-0.70

zeb

-0.70

noon

-0.69

ellow

-0.69

dropping

-0.65

Portland

-0.62

cit

-0.62

church

-0.62

POSITIVE LOGITS

 onstage

0.75

ento

0.65

 Cancel

0.64

ãĥķ

0.63

 ..........

0.62

 Horowitz

0.62

 Emirates

0.61

Ðº

0.60

 Azerbaijan

0.60

 captcha

0.60

Activations Density 0.014%

phrases related to errors or mistakes

phrases related to mistakes or failures

No Comments

No Known Activations