INDEX

Explanations

sequences related to programming or technical instructions

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

Technical code and programming terms

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

脚注の使い方

-0.52

.*")]

-0.49

 bruja

-0.42

posium

-0.42

iſen

-0.40

kesha

-0.40

Hozzáférés

-0.40

 publiques

-0.40

ürger

-0.40

zame

-0.39

POSITIVE LOGITS

Jeografia

0.66

GEBURTSDATUM

0.63

parsedMessage

0.61

styleType

0.57

 Italijanski

0.56

 ویکی‌پدی

0.55

tangentMode

0.54

featureID

0.51

 Савезне

0.51

SharedCtor

0.51

Activations Density 0.152%

sequences related to programming or technical instructions

Technical code and programming terms

No Comments

No Known Activations