INDEX

Explanations

references to specific subjects or entities in relation to their characteristics or actions

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-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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Comté

-0.43

OGND

-0.41

 scarce

-0.37

 EconPapers

-0.35

 referenties

-0.35

prefixer

-0.35

WaitGroup

-0.35

 Bethany

-0.34

 lieux

-0.34

 Unterricht

-0.34

POSITIVE LOGITS

contentLoaded

0.58

 whereas

0.54

Whereas

0.52

 Whereas

0.51

 Sementara

0.49

èdia

0.49

EndInit

0.48

tetten

0.47

AndroidJUnit

0.46

 iconTwitter

0.45

Activations Density 0.024%

references to specific subjects or entities in relation to their characteristics or actions

No Comments

No Known Activations