INDEX

Explanations

This neuron seems to identify elements of code and mathematical problems, particularly focusing on keywords like "command", "highest," "UInt," mathematical operators, and array-like structures.

oai_token-act-pair · gemini-2.0-flash

code

np_max-act · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_8/width_16k/average_l0_52

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.8.ln2.hook_normalized

Architecture

jumprelu_transcoder

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

DoubleQuotes

-0.50

samo

-0.46

Dei

-0.44

 actu

-0.43

subtype

-0.42

 المعيارى

-0.42

hup

-0.41

uacute

-0.40

同年

-0.40

Ύ

-0.39

POSITIVE LOGITS

fjspx

0.74

richTextPanel

0.72

 kaarangay

0.70

 AssemblyCulture

0.69

 externi

0.66

Personendaten

0.66

mybatisplus

0.63

 quæ

0.62

 Jérusalem

0.61

principalColumn

0.60

Activations Density 18.400%

This neuron seems to identify elements of code and mathematical problems, particularly focusing on keywords like "command", "highest," "UInt," mathematical operators, and array-like structures.

code

No Comments

No Known Activations