INDEX

Explanations

repeated mentions of the term "magic."

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

magic and magical things

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

<tr>

-0.40

 voet

-0.39

an

-0.36

 пунктов

-0.34

 stool

-0.33

 puesto

-0.32

diye

-0.32

 مراجع

-0.32

 been

-0.32

 persoane

-0.32

POSITIVE LOGITS

 Magic

1.17

 MAGIC

1.15

Magic

1.09

 magic

1.09

magic

1.05

MAGIC

0.99

 magia

0.89

 magique

0.82

 mág

0.79

magical

0.78

Activations Density 0.010%

repeated mentions of the term "magic."

magic and magical things

No Comments

No Known Activations