INDEX

Explanations

structured lists and instructions

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

numbered list item starts

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

 Administrativna

-0.77

 kaarangay

-0.76

ंदीखरीदारी

-0.75

TypedDataSet

-0.68

цездатний

-0.67

 utafitiHapana

-0.65

出版年

-0.61

Хьажоргаш

-0.58

EndProject

-0.58

 мәкал

-0.57

POSITIVE LOGITS

Hero

0.39

 viewDidLoad

0.34

 Hero

0.33

Hist

0.32

 hero

0.32

 ménage

0.30

hero

0.29

Adelaide

0.28

 Adelaide

0.28

anco

0.28

Activations Density 0.023%

structured lists and instructions

numbered list item starts

No Comments

No Known Activations

structured lists and instructions

numbered list item starts

No Comments

No Known Activations