INDEX

Explanations

specific terms and identifiers related to technical or scientific data

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

figures, propositions, or tables

np_acts-logits-general · gemini-2.0-flash

mathematical or scientific citations

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 intérêt

-0.36

 distributions

-0.35

 élevé

-0.35

なる

-0.34

 residencia

-0.34

 benne

-0.33

 interest

-0.33

 surprise

-0.32

 intéressante

-0.32

 çeşit

-0.32

POSITIVE LOGITS

Datuak

0.85

Portail

0.76

出版年

0.71

SharedCtor

0.70

Jeografia

0.69

 nakalista

0.68

Portale

0.68

parsedMessage

0.66

最快更新

0.66

 дописавши

0.65

Activations Density 0.616%

specific terms and identifiers related to technical or scientific data

figures, propositions, or tables

mathematical or scientific citations

No Comments

No Known Activations