INDEX

Explanations

references to legal and courtroom proceedings, particularly involving murder cases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ViewFeatures

-0.46

WebElementEntity

-0.44

 direcciones

-0.42

 educativas

-0.41

 väli

-0.39

NameInMap

-0.39

httphttps

-0.38

 fine

-0.38

 telefónica

-0.37

 inalámbrica

-0.37

POSITIVE LOGITS

 murder

0.65

 murders

0.62

 murdered

0.62

murder

0.61

 killings

0.58

AsUp

0.57

 homicide

0.56

 slain

0.56

Murder

0.56

 Murder

0.55

Activations Density 0.340%

references to legal and courtroom proceedings, particularly involving murder cases

No Comments

No Known Activations