INDEX

Explanations

references to living spaces, structures, or properties associated with community features

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

multiple items counted

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

Millan

-0.44

 препратки

-0.44

MockBean

-0.43

IContainer

-0.43

CrossRef

-0.42

bip

-0.41

horn

-0.41

″]

-0.40

 Monterey

-0.40

⎠

-0.39

POSITIVE LOGITS

ChildScrollView

0.56

NewUrlParser

0.56

 multiple

0.47

multiple

0.47

⤹

0.46

SharedDtor

0.45

Multiple

0.44

 MULTIPLE

0.42

 múltiples

0.42

 múltiple

0.42

Activations Density 0.081%

references to living spaces, structures, or properties associated with community features

multiple items counted

No Comments

No Known Activations