INDEX

Explanations

phrases related to arrangement or alignment in a specific order

oai_token-act-pair · gpt-3.5-turbo

phrases that include the word "lined."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ENA

-0.68

prov

-0.66

 Consumption

-0.65

 governing

-0.63

authent

-0.60

 mischief

-0.59

 respecting

-0.59

 revoke

-0.58

eus

-0.58

 Admin

-0.58

POSITIVE LOGITS

 lined

3.77

lined

1.55

 dotted

1.48

 lining

1.47

 littered

1.25

 piled

1.25

 covered

1.24

 circled

1.22

 jammed

1.20

 coated

1.19

Activations Density 0.006%

phrases related to arrangement or alignment in a specific order

phrases that include the word "lined."

No Comments

No Known Activations