INDEX

Explanations

mathematical operations and expressions

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

Code, often with hexadecimal

np_acts-logits-general · gemini-2.0-flash

comma followed by numbers or names

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

TypedDataSet

-0.80

يكب

-0.71

 ProtoMessage

-0.69

WaitGroup

-0.68

ValueGeneration

-0.65

richTextPanel

-0.56

EndContext

-0.56

UnknownFieldSet

-0.55

Населення

-0.55

 مرئيه

-0.54

POSITIVE LOGITS

SuppressLint

0.34

httphttps

0.32

 Texans

0.30

fxml

0.30

 Lehr

0.29

Kjelder

0.29

TargetApi

0.29

InputModule

0.28

الدراسه

0.28

 Reise

0.28

Activations Density 1.403%

mathematical operations and expressions

Code, often with hexadecimal

comma followed by numbers or names

No Comments

No Known Activations