INDEX

Explanations

phrases related to specific entities or concepts, possibly involving conflicts or controversies

oai_token-act-pair · gpt-3.5-turbo

mentions of a specific character or entity symbolized by the character "Ļ"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 disadvant

-0.75

 misunder

-0.72

 mathemat

-0.66

 contrace

-0.65

 seiz

-0.65

 condem

-0.64

 Palestin

-0.64

merce

-0.63

 regulation

-0.63

ozy

-0.63

POSITIVE LOGITS

ï¸ı

1.48

ï¸

0.96

âĹ

0.94

0.92

Balt

0.86

âĸº

0.84

gypt

0.83

Â¯Â¯

0.83

âĪ

0.83

âĻ

0.82

Activations Density 0.449%

phrases related to specific entities or concepts, possibly involving conflicts or controversies

mentions of a specific character or entity symbolized by the character "Ļ"

No Comments

No Known Activations