INDEX

Explanations

informational phrases and announcements

oai_token-act-pair · gpt-3.5-turbo

references to sources of information and requests for further details

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

artifacts

-0.74

 neighb

-0.70

lihood

-0.68

 prophes

-0.62

igham

-0.60

inventoryQuantity

-0.59

opped

-0.59

 metic

-0.59

 misunder

-0.59

eatures

-0.59

POSITIVE LOGITS

 about

1.18

 ABOUT

1.08

 regarding

1.06

About

1.01

about

0.95

 About

0.94

 concerning

0.85

 pertaining

0.85

REG

0.83

 âĢº

0.83

Activations Density 0.068%

informational phrases and announcements

references to sources of information and requests for further details

No Comments

No Known Activations