INDEX

Explanations

references to professional roles and personal credentials

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

Comes before attribution phrases (e.g., said, testified)

np_acts-logits-general · gemini-2.0-flash

end of sentenceend of clauselist separator

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

 autorytatywna

-0.63

Diweddarwch

-0.60

 nahilalakip

-0.58

WindowConstants

-0.56

 InputDecoration

-0.53

 OMITBAD

-0.52

 betweenstory

-0.52

TagMode

-0.49

complexContent

-0.49

">—

-0.49

POSITIVE LOGITS

quién

0.38



0.37

 mówią

0.37

Pontos

0.37

who

0.37

 profesor

0.36

leader

0.36

 esquerda

0.36

的说道

0.35

 advised

0.35

Activations Density 0.244%

references to professional roles and personal credentials

Comes before attribution phrases (e.g., said, testified)

end of sentenceend of clauselist separator

No Comments

No Known Activations