INDEX

Explanations

phrases related to embedding content

oai_token-act-pair · gpt-3.5-turbo

references to embedding content, particularly in digital formats

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

»Ĵ

-0.80

externalActionCode

-0.77

othal

-0.77

ierra

-0.70

Pis

-0.69

asonable

-0.65

 Bureau

-0.64

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

-0.63

 hust

-0.62

ģĸ

-0.61

POSITIVE LOGITS

embed

1.37

ding

1.35

ded

1.05

 embed

1.02

iframe

0.99

rities

0.85

dict

0.83

mers

0.81

gon

0.80

ming

0.79

Activations Density 0.019%

phrases related to embedding content

references to embedding content, particularly in digital formats

No Comments

No Known Activations