INDEX

Explanations

instances of formal statements or admissions of wrongdoing

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

RTSC

-0.44

 keker

-0.42

 koncep

-0.41

 Umfang

-0.41

 conformidad

-0.41

 conçu

-0.41

 conç

-0.40

 avanzada

-0.40

 convivencia

-0.40

 znac

-0.40

POSITIVE LOGITS

 marmo

0.50

 resourceCulture

0.50

Clik

0.48

دانشنامهٔ

0.48

GEBURTS

0.47

cesis

0.47

SharedDtor

0.47

 rebo

0.46

 pats

0.45

 Vodka

0.45

Activations Density 0.141%

instances of formal statements or admissions of wrongdoing

No Comments

No Known Activations