INDEX

Explanations

code structures, particularly class definitions and their access modifiers

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

javabean and c++ code fragments

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

 jScrollPane

-0.54

WriteTagHelper

-0.53

Autoritní

-0.47

 Roskov

-0.47

 kasarigan

-0.47

complexContent

-0.47

Aholisi

-0.43

 ModelExpression

-0.43

routeProvider

-0.43

 تضيفلها

-0.42

POSITIVE LOGITS

UnusedPrivate

0.54

SourceChecksum

0.50

✷

0.49

Tail

0.48

issão

0.46

Bes

0.45

 Kolon

0.45

Vehicle

0.44

Best

0.43

 المعيارى

0.43

Activations Density 0.017%

code structures, particularly class definitions and their access modifiers

javabean and c++ code fragments

No Comments

No Known Activations