INDEX

Explanations

references to programs and initiatives

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

programme

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

 verrez

-0.62

IBOutlet

-0.62

 fumée

-0.56

 fédé

-0.54

addContainerGap

-0.54

Hik

-0.54

 Púb

-0.51

存于互联网档案馆

-0.50

 PagesJaunes

-0.50

 мәкал

-0.49

POSITIVE LOGITS

 programme

0.85

 Programme

0.83

programme

0.81

Programme

0.80

 PROGRAMME

0.79

 programmes

0.61

ambique

0.60

 Programmes

0.59

rophore

0.55

 Europa

0.54

Activations Density 0.005%

references to programs and initiatives

programme

No Comments

No Known Activations