INDEX

Explanations

numeric values related to legal or procedural contexts

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

numbers followed by section symbols

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

Negative Logits

 تضيفلها

-0.50

<tfoot>

-0.41

रिक

-0.39

INKS

-0.38

illette

-0.38

ent

-0.36

armed

-0.35

 colspan

-0.35

made

-0.35

ärs

-0.34

POSITIVE LOGITS

AccessorTable

0.76

XmlAccessType

0.73

 MessageBoxIcon

0.69

Aholisi

0.68

 autorytatywna

0.59

$_['

0.59

NameInMap

0.58

httphttps

0.56

RectangleBorder

0.56

SharedDtor

0.55

Activations Density 0.058%

numeric values related to legal or procedural contexts

numbers followed by section symbols

No Known Activations