INDEX

Explanations

references to firefighting and wildfire events

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

wildfire evacuation

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

dańsk

-0.31

 toim

-0.31

ver

-0.31

Olvid

-0.30

 occur

-0.29

 ljud

-0.29

 Lozano

-0.29

 displays

-0.28

display

-0.28

gucig

-0.28

POSITIVE LOGITS

 wildfire

0.69

 wildfires

0.67

RectangleBorder

0.65

 firefighting

0.64

SBATCH

0.62

parsedMessage

0.61

 firefighters

0.61

 disambiguazione

0.60

󠁴

0.59

 firefighter

0.59

Activations Density 0.022%

references to firefighting and wildfire events

wildfire evacuation

No Comments

No Known Activations