INDEX

Explanations

negative or conditional expressions

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

starting with `!`

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

يكب

-0.44

 setImage

-0.42

 Italijani

-0.35

 бере

-0.34

 hopefully

-0.34

ollectionView

-0.33

 informée

-0.33

 prochaines

-0.32

 rör

-0.31

Бере

-0.31

POSITIVE LOGITS

IsContent

0.90

(!__

0.84

(!$

0.70

(!$

0.65

(!

0.60

Personendaten

0.59

 initComponents

0.57

HasIndex

0.54

niająca

0.54

">//

0.53

Activations Density 0.008%

negative or conditional expressions

starting with `!`

No Comments

No Known Activations