INDEX

Explanations

phrases related to official spokespersons

oai_token-act-pair · gpt-3.5-turbo

the word "spokesman" and variations of it

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

achable

-0.74

ī

-0.65

 Forestry

-0.63

achi

-0.61

iller

-0.61

Yaz

-0.60

 âĪ¼

-0.60

wx

-0.60

ï¸ı

-0.59

osition

-0.59

POSITIVE LOGITS

Interstitial

0.78

tein

0.75

 skelet

0.73

wings

0.70

swick

0.70

banks

0.66

 pregn

0.65

kefeller

0.64

 ILCS

0.64

taboola

0.63

Activations Density 0.000%

phrases related to official spokespersons

the word "spokesman" and variations of it

No Comments

No Known Activations