INDEX

Explanations

places or names related to a specific city, specifically Indianapolis

oai_token-act-pair · gpt-3.5-turbo

references to the city of Indianapolis

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

wolf

-0.73

wolves

-0.70

Mu

-0.70

 spider

-0.66

ozy

-0.65

byn

-0.65

 wolf

-0.64

 wolves

-0.64

 diction

-0.64

 biting

-0.62

POSITIVE LOGITS

 Indianapolis

3.74

 Indy

2.23

Indian

2.03

 Colts

1.88

 Indiana

1.79

Indiana

1.75

 Honolulu

1.68

 Louisville

1.66

 Cincinnati

1.65

 Jacksonville

1.60

Activations Density 0.017%

places or names related to a specific city, specifically Indianapolis

references to the city of Indianapolis

No Comments

No Known Activations