INDEX

Explanations

words related to breaking news alerts and special reports

oai_token-act-pair · gpt-3.5-turbo

phrases related to breaking news updates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Supervisor

-0.75

 Mechan

-0.68

 Mutual

-0.68

 Teach

-0.66

 Mistress

-0.66

 Rampage

-0.66

terness

-0.65

 Barber

-0.64

 Catal

-0.63

 Laboratories

-0.62

POSITIVE LOGITS

 news

0.85

 NEWS

0.80

news

0.78

 emergencies

0.70

 inquiries

0.70

Trend

0.70

 stories

0.69

archives

0.67

 investig

0.64

new

0.64

Activations Density 0.024%

words related to breaking news alerts and special reports

phrases related to breaking news updates

No Comments

No Known Activations