INDEX

Explanations

the word "rampant" or similar variations of it

oai_token-act-pair · gpt-3.5-turbo

the prevalence of negative conditions or behaviors 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 @ 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

tein

-0.88

lder

-0.70

 therapists

-0.65

cknow

-0.65

izons

-0.63

psey

-0.63

ney

-0.62

abies

-0.60

rams

-0.60

nam

-0.60

POSITIVE LOGITS

 rampant

1.28

 prevalent

0.71

 speculation

0.70

PDATE

0.70

interstitial

0.68

 widespread

0.67

odox

0.67

 criminality

0.67

 unchecked

0.67

 pervasive

0.65

Activations Density 0.010%

the word "rampant" or similar variations of it

the prevalence of negative conditions or behaviors in various contexts

No Comments

No Known Activations