INDEX

Explanations

instances of structured data or format indicators in text

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

start of user turn

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

Personendaten

-0.53

Rhestr

-0.46

 detailsLots

-0.45

፩

-0.45

ToBounds

-0.45

}{*}{}

-0.44

 Chwiliwch

-0.44

CardHeader

-0.43

 Italijani

-0.42

 utafitiHapana

-0.42

POSITIVE LOGITS

IUrlHelper

0.43

RefNanny

0.42

########.

0.40

хьтан

0.39

jsxFileName

0.39

الحياه

0.37

 defaultstate

0.36

élaboration

0.35

AutoModerator

0.35

BeginInit

0.35

Activations Density 0.000%

instances of structured data or format indicators in text

start of user turn

No Comments

No Known Activations