INDEX

Explanations

sequences of characters or special symbols, possibly related to code or formatting

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

rungsseite

-0.66

:✨

-0.63

 nahilalakip

-0.62

ChildScrollView

-0.60

 noqa

-0.60

 ſtate

-0.59

ſchaft

-0.57

 houſe

-0.57

 greateſt

-0.57

POSITIVE LOGITS

Espagne

0.39

Italie

0.34

 reciclado

0.33

 otomatig

0.32

Irlande

0.32

Empereur

0.32

Allemagne

0.31

 Sénat

0.31

 esprits

0.30

BlockPos

0.30

Activations Density 0.359%

sequences of characters or special symbols, possibly related to code or formatting

No Comments

No Known Activations