INDEX

Explanations

structure and organization in written content

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

category, references, notes

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

amblea

-0.40

 gesteld

-0.37

 betrekking

-0.37

rungsseite

-0.36

 Italijanski

-0.36

 hyö

-0.35

 gdyż

-0.35

 bepaalde

-0.35

 Meksiku

-0.35

daction

-0.35

POSITIVE LOGITS

RectangleBorder

0.68

Aiheesta

0.68

aarrggbb

0.60

 ویکی‌پدیا

0.56

OGND

0.54

 referrerpolicy

0.50

 initComponents

0.49

 تضيفلها

0.49

 cornice

0.47

بوابة

0.47

Activations Density 0.010%

structure and organization in written content

category, references, notes

No Known Activations