INDEX

Explanations

abbreviations or acronyms in scientific or scholarly contexts

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

initials followed by periods

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

 lehe

-0.43

 seragam

-0.43

«

-0.42

丘

-0.41

 murni

-0.41

 pure

-0.40

-0.39

主

-0.39

 handle

-0.38

 ruote

-0.38

POSITIVE LOGITS

 Roskov

0.61

 EconPapers

0.61

 للاسماء

0.58

Билгалдахарш

0.57

 Bourgoin

0.54

 defaultstate

0.54

出版年

0.52

 MainAxisSize

0.52

󠁣

0.50

ViewFeatures

0.50

Activations Density 0.000%

abbreviations or acronyms in scientific or scholarly contexts

initials followed by periods

No Comments

No Known Activations