INDEX

Explanations

abbreviations related to professional titles

oai_token-act-pair · gpt-3.5-turbo

references to names and titles, particularly those related to media or notable individuals

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

aucus

-0.72

fman

-0.71

cius

-0.71

holders

-0.70

milo

-0.69

clusive

-0.64

uity

-0.64

iture

-0.64

utra

-0.63

ueless

-0.62

POSITIVE LOGITS

ournal

0.87

 Edgar

0.87

Lo

0.83

Crew

0.77

 Sawyer

0.75

League

0.75

 Marriott

0.74

 Jonah

0.73

 Styles

0.73

Ax

0.72

Activations Density 0.051%

abbreviations related to professional titles

references to names and titles, particularly those related to media or notable individuals

No Comments

No Known Activations