INDEX

Explanations

specific names, brands, or entities within texts

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

Names or brands followed by their components

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

Negative Logits

 protoimpl

-0.52

ViewFeatures

-0.48

featureID

-0.43

SpringBootTest

-0.41

[*]

-0.40

enumii

-0.40

 ProtoMessage

-0.38

┻

-0.38

AnimationsModule

-0.38

/***/

-0.38

POSITIVE LOGITS

setupUi

0.46

 Paglinawan

0.46

 noDo

0.45

Дереккөздер

0.45

 Drapeau

0.44

ishna

0.43

HasColumnName

0.43

Бахар

0.42

 pleaſure

0.42

Jewel

0.40

Activations Density 0.138%

specific names, brands, or entities within texts

Names or brands followed by their components

No Known Activations