INDEX

Explanations

phrases indicating trust and reliability in services or experiences

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

confidence and safety

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

 AssemblyTitle

-0.73

 للاسماء

-0.69

InitVars

-0.63

 utafitiHapana

-0.63

FunctionFlags

-0.62

 Мексичка

-0.61

GraphicsUnit

-0.61

 houſe

-0.59

 autorytatywna

-0.59

 initComponents

-0.59

POSITIVE LOGITS

safe

0.57

 safe

0.49

安全的

0.39

Safe

0.38

 safety

0.37

 sicheren

0.37

safety

0.36

 Safe

0.35

安全

0.34

 safely

0.34

Activations Density 0.011%

phrases indicating trust and reliability in services or experiences

confidence and safety

No Comments

No Known Activations