INDEX

Explanations

phrases related to protection or safety

oai_token-act-pair · gpt-3.5-turbo

keywords related to protection and safety

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 neoc

-0.75

 nodd

-0.74

ickson

-0.74

 throat

-0.72

 cinem

-0.71

eyed

-0.70

geries

-0.70

ikawa

-0.70

 carn

-0.69

 hairst

-0.68

POSITIVE LOGITS

 Protect

2.91

Protect

2.76

 Safe

2.00

Safe

1.82

protect

1.72

 Help

1.50

 Secure

1.49

 Prevent

1.47

Free

1.39

 Protector

1.37

Activations Density 0.041%

phrases related to protection or safety

keywords related to protection and safety

No Comments

No Known Activations