INDEX

Explanations

elements and tags related to HTML or template structure

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

user inputs or file paths

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

########.

-0.65

ագրություններ

-0.62

tagHelperRunner

-0.62

+#+

-0.61

 autorytatywna

-0.60

httphttps

-0.58

ſſung

-0.54

Хьажоргаш

-0.54

HasAnnotation

-0.54

 विश्वसनीयता

-0.54

POSITIVE LOGITS

richTextPanel

0.38

ثل

0.32

toMatchSnapshot

0.32

around

0.29

 Tempel

0.28

WebElementEntity

0.28

 Adder

0.27

ジング

0.27

 Saal

0.27

 mantle

0.27

Activations Density 0.007%

elements and tags related to HTML or template structure

user inputs or file paths

No Comments

No Known Activations