INDEX

Explanations

mentions of airplanes or plane-related incidents

oai_token-act-pair · gpt-3.5-turbo

references to aircraft or aviation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

laus

-1.03

PUT

-0.85

essee

-0.78

UGC

-0.75

OFF

-0.73

 Cumber

-0.72

optional

-0.72

FINE

-0.70

lishes

-0.70

é»Ĵ

-0.68

POSITIVE LOGITS

 planes

1.08

 plane

1.08

plane

1.08

 airplanes

1.06

planes

1.01

 airliner

1.01

 Planes

0.99

 hangar

0.95

 carrier

0.94

 prope

0.93

Activations Density 0.019%

mentions of airplanes or plane-related incidents

references to aircraft or aviation

No Comments

No Known Activations