INDEX

Explanations

mentions of names or proper nouns

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

proper nouns, places, brands

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

 autorytatywna

-0.69

Autoritní

-0.56

 ***!

-0.51

 nakalista

-0.50

 שוליים

-0.50

 الرياضيه

-0.47

__':

-0.46

Personendaten

-0.46

posedge

-0.46

 tartalomajánló

-0.45

POSITIVE LOGITS

0.54

Gön

0.54

 rispar

0.52

 riscal

0.49

 bezpiecze

0.48

 zupeł

0.48

 îmb

0.48

 zelve

0.48

 szczegó

0.46

Reyn

0.46

Activations Density 0.021%

mentions of names or proper nouns

proper nouns, places, brands

No Comments

No Known Activations