INDEX

Explanations

references to injuries and wounds

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 40-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_40/width_16k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.40.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

 ब्रेकडाउन

-0.47

drawSprites

-0.47

Hurt

-0.42

endix

-0.41

 malheur

-0.41

Билгалдахарш

-0.40

sana

-0.40

 Hurt

-0.40

 malade

-0.39

accident

-0.39

POSITIVE LOGITS

 cuts

0.73

 abra

0.67

 lacer

0.65

 burns

0.62

 contu

0.61

 superficial

0.56

 fractures

0.54

 internal

0.54

 Cuts

0.52

 slashes

0.52

Activations Density 0.504%

references to injuries and wounds

No Comments

No Known Activations