INDEX

Explanations

quoted statements or direct speech within the text

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

end of quotes or sentences

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

Red

-0.47

-0.46

 Central

-0.42

 travelers

-0.39

 hard

-0.38

red

-0.38

Inc

-0.38

red

-0.38

 residents

-0.38

-0.37

POSITIVE LOGITS

 Италијани

0.75

ViewFeatures

0.62

featureID

0.58

writeField

0.57

parsedMessage

0.56

ftagPool

0.55

 незавершена

0.53

Rhestr

0.53

nehin

0.53

 >=",

0.52

Activations Density 0.075%

quoted statements or direct speech within the text

end of quotes or sentences

No Comments

No Known Activations