INDEX

Explanations

tokens that represent common programming, technical terms, or identifiers typically associated with code and programming contexts

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

fire extinguisher; trade; raw; son; NSError

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

 Sosten

-0.37

nasium

-0.32

Obrázky

-0.30

 Schließlich

-0.28

 inoxydable

-0.27

 schließlich

-0.27

CppMethod

-0.27

 sobres

-0.27

帖最后由

-0.27

sizePolicy

-0.27

POSITIVE LOGITS

expandindo

0.72

LookAnd

0.69



0.65

Autoritní

0.62

RectangleBorder

0.58

 Demp

0.55

SerializedName

0.55

 Taktlose

0.54

rungsseite

0.51

Mund

0.51

Activations Density 0.010%

tokens that represent common programming, technical terms, or identifiers typically associated with code and programming contexts

fire extinguisher; trade; raw; son; NSError

No Comments

No Known Activations