INDEX

Explanations

phrases addressing permissions and user instructions

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

You can specify

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

 يتيمه

-0.77

 betweenstory

-0.77

RegressionTest

-0.77

UserScript

-0.72

fromnode

-0.67

 okuyayım

-0.66

:✨

-0.63

 Administrativna

-0.62

 autorytatywna

-0.62

IntoConstraints

-0.62

POSITIVE LOGITS

Становништво

0.36

 econô

0.31

tagHelperRunner

0.31

 facteurs

0.30

جستارهای

0.29

 AssemblyCulture

0.29

Glej

0.28

 productivo

0.27

 factores

0.27

 privil

0.26

Activations Density 0.000%

phrases addressing permissions and user instructions

You can specify

No Comments

No Known Activations