INDEX

Explanations

identifiers and attributes related to structured data formats

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

command definitions

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

TypedDataSet

-0.81

tagHelperRunner

-0.79

 メンテナ

-0.77



-0.73

ロウィン

-0.73

ſſung

-0.72

<unused7>

-0.71

󠁁

-0.71

$_-

-0.71

<unused68>

-0.71

POSITIVE LOGITS

 Wisata

0.30

 Zusammensetzung

0.29

jsxFileName

0.27

 saludo

0.27

 Oberfläche

0.27

 provisoire

0.26

 ucapnya

0.25

 Lingkungan

0.24

 Penyakit

0.24

BrowserRouter

0.23

Activations Density 0.001%

identifiers and attributes related to structured data formats

command definitions

No Comments

No Known Activations