INDEX

Explanations

phrases indicating various degrees or levels of something

oai_token-act-pair · gpt-3.5-turbo

varying degrees or qualities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

idays

-0.75

iddler

-0.75

wine

-0.70

idav

-0.70

letters

-0.66

feeding

-0.65

 Freddie

-0.65

INTON

-0.64

ischer

-0.64

pill

-0.63

POSITIVE LOGITS

 popularity

1.29

 prominence

1.27

 strength

1.26

 fame

1.25

 complexity

1.24

 sophistication

1.22

 potency

1.22

 notoriety

1.21

 usefulness

1.21

 versatility

1.19

Activations Density 0.579%

phrases indicating various degrees or levels of something

varying degrees or qualities

No Comments

No Known Activations