INDEX

Explanations

articles and quantifiers in the text

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

words following articles and pronouns

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ValueStyle

-0.50

utilisons

-0.40

 cavalli

-0.36

 TAMBÉM

-0.35

 zomer

-0.34

 ménage

-0.34

Autoritní

-0.33

DispatchToProps

-0.33

 pets

-0.32

 ivelany

-0.31

POSITIVE LOGITS

featureID

0.59

ontale

0.49

kijken

0.49

何より

0.48

inWeight

0.48

 reaſon

0.48

 rasanya

0.48

principalTable

0.47

 menyebut

0.47

Especially

0.46

Activations Density 0.121%

articles and quantifiers in the text

words following articles and pronouns

No Comments

No Known Activations