INDEX

Explanations

programming-related keywords and directives

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

CPE

-0.39

ArgsConstructor

-0.38

ModelMap

-0.37

Ắ

-0.36

CPE

-0.35

Jia

-0.35

pcb

-0.35

 wastewater

-0.34

 wong

-0.34

koz

-0.33

POSITIVE LOGITS

 disambiguazione

0.56

 تضيفلها

0.54

 autorytatywna

0.52

Diweddarwch

0.52

Hochspringen

0.52

endregion

0.50

 ujednoznacz

0.50

帖最后由

0.49

 himo

0.49

 Савезне

0.48

Activations Density 0.059%

programming-related keywords and directives

No Comments

No Known Activations