INDEX

Explanations

phrases related to specific locations (like cities or areas)

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

Data Type

torch.float32

Hook Point

blocks.8.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

-1.05

thood

-0.80

letter

-0.72

VALUE

-0.70

ebook

-0.69

ogun

-0.69

arial

-0.69

onym

-0.68

istration

-0.67

rahim

-0.66

POSITIVE LOGITS

 County

1.30

 City

1.28

 Heights

1.19

 Township

1.11

 Municipal

1.10

 Airport

1.08

 suburbs

1.08

 Beach

1.07

 Hills

1.04

 Avenue

1.04

Activations Density 2.652%

phrases related to specific locations (like cities or areas)

No Comments

No Known Activations