INDEX

Explanations

code-related keywords and expressions in programming languages

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

math and code constructs

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

MessageTagHelper

-0.38

出版年

-0.35

 prince

-0.33

 ActionListener

-0.32

 Prince

-0.31

mainAxisSize

-0.30

setScene

-0.29

 exceed

-0.29

Vidite

-0.29

Données

-0.28

POSITIVE LOGITS

brada

0.54

 InputDecoration

0.51

씩

0.50

webElementXpaths

0.49

SequentialGroup

0.49

 ویکی‌پدیای

0.48

lafen

0.48

 Chwiliwch

0.48

 AssemblyCulture

0.47

djangoproject

0.46

Activations Density 0.130%

code-related keywords and expressions in programming languages

math and code constructs

No Comments

No Known Activations