INDEX

Explanations

technical and programming terms relevant to firmware and hardware functionality

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

Marlin manufacturing safety

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

twimg

-0.46

 للاسماء

-0.44

didSet

-0.44

 arşivlendi

-0.43

basicConfig

-0.42

addGap

-0.41

 Paglinawan

-0.40

 esercit

-0.40

 riusc

-0.40

TypedDataSet

-0.39

POSITIVE LOGITS

 Manufacturers

0.49

 Safely

0.46

AndEndTag

0.45

 safely

0.45

 Safety

0.44

 manufacturers

0.41

 Barber

0.40

farwyddwr

0.40

 plan

0.39

 manufacturer

0.39

Activations Density 0.022%

technical and programming terms relevant to firmware and hardware functionality

Marlin manufacturing safety

No Comments

No Known Activations