INDEX

Explanations

references to specific news channels and their characteristics

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

invented

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

principalTable

-0.62

 StatefulWidget

-0.58

 Vikipedi

-0.54

endphp

-0.52

AsUp

-0.48

 informée

-0.47

WriteBarrier

-0.47

ंदीखरीदारी

-0.47

 comuniques

-0.47

 ویکی‌آمباردا

-0.47

POSITIVE LOGITS

Vidite

0.41

 lenker

0.34

 utafitiHapana

0.33

]")]

0.32

OGND

0.32

 بندی

0.31

 azért

0.31

 Ще

0.31

könyv

0.31

ReusableCell

0.30

Activations Density 0.027%

references to specific news channels and their characteristics

invented

No Comments

No Known Activations