INDEX

Explanations

code-related syntax elements and structures

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

Code snippets containing brackets and symbols

np_acts-logits-general · gemini-2.0-flash

code structures around assignments

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.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

GEBURTS

-0.63

 мәкал

-0.54

 autorytatywna

-0.52

=$?

-0.50

Diwedd

-0.50

MarshalTo

-0.50

 <>",

-0.48

endphp

-0.48

 /\.(

-0.47

});*/

-0.46

POSITIVE LOGITS

limard

0.40

الحياه

0.40

 also

0.40

nowu

0.39

 defaultstate

0.39

kuuta

0.38

 ilang

0.38

enschappelijke

0.37

Geografie

0.37

 velvet

0.36

Activations Density 0.410%

code-related syntax elements and structures

Code snippets containing brackets and symbols

code structures around assignments

No Comments

No Known Activations