INDEX

Explanations

references to authority and control

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 41-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_41/width_16k/average_l0_113

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.41.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

thâu

-0.54

 ब्रेकडाउन

-0.49

expandindo

-0.49

 kasarigan

-0.45

 Италијани

-0.45

InitVars

-0.43

 Italijanski

-0.43

żeli

-0.41

الإنجليزية

-0.39

CrossRef

-0.38

POSITIVE LOGITS

 obliquely

0.58

المشاركات

0.49

 inlaid

0.48

 stained

0.46

 nailed

0.45

 embedded

0.44

 tightly

0.43

 Walkover

0.43

 slanted

0.42

 tied

0.41

Activations Density 0.084%

references to authority and control

No Comments

No Known Activations