INDEX

Explanations

programming-related terms, specifically focusing on styles and classes often used in document formatting or programming frameworks

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

style definitions

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

 محفوظة

-0.55

 pouvoit

-0.52

 auroit

-0.48

 betweenstory

-0.48

月号

-0.43

\{\\

-0.43

 resourceCulture

-0.43

pun

-0.43

ambilan

-0.43

tableFuture

-0.42

POSITIVE LOGITS

style

0.61

WebServlet

0.56

openzeppelin

0.48

Билгалдахарш

0.47

CppMethod

0.47

 ErrIntOverflow

0.47

Lähteet

0.47

usepackage

0.45

pushFollow

0.44

GeneratedValue

0.44

Activations Density 0.000%

programming-related terms, specifically focusing on styles and classes often used in document formatting or programming frameworks

style definitions

No Comments

No Known Activations