INDEX

Explanations

terms and structures related to web applications or product suggestion components

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

string 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

Билгалдахарш

-0.39

saraba

-0.36

 realize

-0.35

 déten

-0.34

ang

-0.34

Lav

-0.34

Javadoc

-0.32

 enfant

-0.31

 Wade

-0.31

 realized

-0.30

POSITIVE LOGITS

MemoryWarning

0.59

出版年

0.59

 betweenstory

0.56

Clik

0.52

 BoxFit

0.51

ItemBackground

0.50

StoryboardSegue

0.49

 Monfieur

0.48

 SwitchCompat

0.48

typeDefinition

0.46

Activations Density 0.003%

terms and structures related to web applications or product suggestion components

string definitions

No Comments

No Known Activations