INDEX

Explanations

phrases related to actions taken without authorization or consent

oai_token-act-pair · gpt-3.5-turbo

instances of unauthorized actions or lack of consent

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

NetMessage

-0.73

esa

-0.73

ãĥĺ

-0.70

 âĢº

-0.68

culture

-0.67

cro

-0.67

cephal

-0.66

phis

-0.66

ahime

-0.65

emb

-0.64

POSITIVE LOGITS

 permission

1.26

 authorization

1.25

 consent

1.18

 informing

1.08

 explicit

1.01

 approval

0.98

 notification

0.97

 Consent

0.95

 parental

0.95

 notice

0.92

Activations Density 0.112%

phrases related to actions taken without authorization or consent

instances of unauthorized actions or lack of consent

No Comments

No Known Activations