INDEX

Explanations

geographical references to regions and borders

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_7/width_16k/average_l0_92

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 pala

-0.47

 Latvian

-0.44

HK

-0.43

pala

-0.43

Reader

-0.42

 Hongkong

-0.42

kt

-0.41

DOM

-0.41

sme

-0.40

omet

-0.40

POSITIVE LOGITS

 north

0.92

 selatan

0.90

 east

0.88

 south

0.84

 west

0.81

 utara

0.81

 northern

0.78

 northward

0.75

north

0.75

 northwest

0.74

Activations Density 0.290%

geographical references to regions and borders

No Comments

No Known Activations