INDEX

Explanations

locations and events related to specific cities or states

oai_token-act-pair · gpt-3.5-turbo

geographic locations and names of places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

xual

-0.94

URI

-0.74

strings

-0.68

pse

-0.67

 deduct

-0.65

 exha

-0.64

Versions

-0.64

 nomine

-0.62

 asymm

-0.62

 labelled

-0.62

POSITIVE LOGITS

 Asheville

1.45

 Spokane

1.43

 Sioux

1.39

 Wilmington

1.39

 Indianapolis

1.38

 Chattanooga

1.37

 Minneapolis

1.37

 Albuquerque

1.37

 Nashville

1.37

 Augusta

1.36

Activations Density 0.193%

locations and events related to specific cities or states

geographic locations and names of places

No Comments

No Known Activations