INDEX

Explanations

instances of punctuation followed by specific programming language structures or keywords

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

SQL and .NET operations

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

Personendaten

-0.92

transQ

-0.74

featureID

-0.71

 surla

-0.68

 betweenstory

-0.67

SequentialGroup

-0.66

SBATCH

-0.64

VersionUID

-0.62

 desmotivaciones

-0.61

AsUp

-0.58

POSITIVE LOGITS

::

0.44

].[

0.40

Ken

0.40

Or

0.39

Linq

0.39

BE

0.38

Core

0.37

Sie

0.37

 Original

0.37

Ặ

0.37

Activations Density 0.011%

instances of punctuation followed by specific programming language structures or keywords

SQL and .NET operations

No Comments

No Known Activations