INDEX

Explanations

cities or locations

oai_token-act-pair · gpt-3.5-turbo

references to specific city names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

vana

-0.80

merce

-0.80

acca

-0.77

arist

-0.75

lyak

-0.72

arial

-0.69

ascript

-0.67

utations

-0.66

ogun

-0.66

quo

-0.66

POSITIVE LOGITS

 Mayor

1.07

 suburb

1.06

 Downtown

1.02

 Metropolitan

1.02

 suburbs

1.02

 skyline

1.01

 Airport

1.01

 City

1.01

 mayoral

1.01

 County

1.00

Activations Density 0.128%

cities or locations

references to specific city names

No Comments

No Known Activations