INDEX

Explanations

code annotations and metadata in programming contexts

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

programming annotations and documentation

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

 betweenstory

-0.52

Scrollbar

-0.41

 volantes

-0.40

 afloat

-0.40

 salvaje

-0.39

 arvio

-0.39

 Bezirks

-0.39

ToAction

-0.38

 asti

-0.37

 asesinado

-0.37

POSITIVE LOGITS

 autorytatywna

0.72

 otomatig

0.68

styleType

0.65

NameInMap

0.57

WireFormatLite

0.57

DataAnnotations

0.56

 IBOutlet

0.54

AnchorStyles

0.52

 Савезне

0.51

ApiModelProperty

0.50

Activations Density 0.001%

code annotations and metadata in programming contexts

programming annotations and documentation

No Comments

No Known Activations