INDEX

Explanations

words and phrases indicating measurements, updates, and variations in data or situations

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

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

 houſe

-0.36

 consequently

-0.35

 successful

-0.34

 striking

-0.33

 result

-0.32

endregion

-0.32

いかがでしょうか

-0.32

tor

-0.32

 Schen

-0.32

 indeed

-0.31

POSITIVE LOGITS

 informée

0.75

principalTable

0.75

 المعيارى

0.73

 Administrativna

0.66

contentLoaded

0.62

fjspx

0.60

ValueStyle

0.60

 виправивши

0.60

Personensuche

0.56

Datuak

0.56

Activations Density 0.094%

words and phrases indicating measurements, updates, and variations in data or situations

No Comments

No Known Activations