INDEX

Explanations

variations of the word "to" and prepositions highlighting relationships or characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 MainAxisSize

-0.52

 uſed

-0.50

 tranſ

-0.49

enumi

-0.48

 HttpHeaders

-0.48

 detención

-0.47

 juſ

-0.46

 anſ

-0.44

Muerte

-0.44

 againſt

-0.43

POSITIVE LOGITS

RegistryLite

0.51

 наконец

0.49

المكان

0.46

 lastly

0.46

 otomatig

0.45

 typelib

0.44

 {}));

0.43

 ivelany

0.43

})*/

0.43

 enfin

0.42

Activations Density 0.876%

variations of the word "to" and prepositions highlighting relationships or characteristics

No Comments

No Known Activations