INDEX

Explanations

phrases related to allegations or accusations of wrongdoing

oai_token-act-pair · gpt-3.5-turbo

references to claims or statements of wrongdoing or misconduct

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

ger

-0.68

xual

-0.67

ilation

-0.66

atu

-0.66

bern

-0.65

ament

-0.65

liv

-0.65

ature

-0.64

focus

-0.63

heses

-0.63

POSITIVE LOGITS

 violated

0.76

 misrepresent

0.75

 allegedly

0.73

 metic

0.73

 infringing

0.72

 contradict

0.72

Buyable

0.71

æ©

0.71

 originated

0.71

 infring

0.71

Activations Density 0.006%

phrases related to allegations or accusations of wrongdoing

references to claims or statements of wrongdoing or misconduct

No Comments

No Known Activations