INDEX

Explanations

names and identities, particularly related to inventors and authors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 39-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_39/width_16k/average_l0_131

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.39.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

 Indones

-0.45

 Indon

-0.42

GTCX

-0.40

 Borneo

-0.39

 chines

-0.39

 Indonesian

-0.38

ſelf

-0.37

Indonesia

-0.37

 Puerto

-0.37

 Indonesia

-0.37

POSITIVE LOGITS

 Armenian

1.29

 Armenia

1.26

 armen

1.12

Armenia

1.09

 Yerevan

1.08

Armen

1.07

 Armenians

1.04

 Armen

1.03

armen

0.86

 Арме

0.85

Activations Density 0.075%

names and identities, particularly related to inventors and authors

No Comments

No Known Activations