INDEX

Explanations

code-related terms and structures, particularly those associated with the end of functions and procedures

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

end of code block or statement

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

TargetException

-0.52

 ویکی‌پدی

-0.52

المناصب

-0.52

sendRedirect

-0.48

 cousins

-0.48

MatInputModule

-0.48

延伸閱讀

-0.47

ReadKey

-0.47

 intptr

-0.47

webtoken

-0.47

POSITIVE LOGITS

InjectAttribute

0.41

afterEach

0.35

AfterEach

0.34

beforeEach

0.34

Prí

0.32

Autowired

0.31

 kriti

0.30

 VIAF

0.30

 prí

0.30

こと

0.29

Activations Density 0.001%

code-related terms and structures, particularly those associated with the end of functions and procedures

end of code block or statement

No Comments

No Known Activations