INDEX

Explanations

references to programming concepts and structures

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

shader code and technical terms

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

kuuta

-0.49

PyExc

-0.44

fjspx

-0.44

ADE

-0.39

 TextAppearance

-0.38

 thin

-0.38

 ني

-0.38

uguetes

-0.38

 jspb

-0.37

xay

-0.36

POSITIVE LOGITS

 évêque

0.51

shadowRadius

0.49

bootstrapcdn

0.48

expandindo

0.47

DispatchToProps

0.46

 AssemblyProduct

0.45

 onCancelled

0.44

0.43

TokenNameDOT

0.42

denominator

0.41

Activations Density 0.357%

references to programming concepts and structures

shader code and technical terms

No Comments

No Known Activations