INDEX

Explanations

numbers and code syntax

np_acts-logits-general · gemini-2.5-flash-lite

The neuron detects numeric literals (integer or floating‐point numbers) in code or text.

oai_token-act-pair · o4-mini Triggered by @jyhe0408

numeric literals and array/matrix dimensions in programming code.

oai_token-act-pair · claude-4-5-sonnet Triggered by @jyhe0408

numeric literals in code, especially array/matrix dimensions, sizes, and indices within shapes, brackets, and function calls

oai_token-act-pair · gpt-5 Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_22/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 until

-1.69

 immediately

-1.56

or

-1.55

 before

-1.53

 this

-1.52

if

-1.52

by

-1.51

 remained

-1.45

 focusing

-1.43

 during

-1.42

POSITIVE LOGITS

 geliefert

1.53

another

1.51

 también

1.39

傒

1.39

 utiliser

1.31

 каждой

1.30

本当に

1.27

 Dinas

1.26

 também

1.24

 uniquement

1.24

Activations Density 0.015%

numbers and code syntax

The neuron detects numeric literals (integer or floating‐point numbers) in code or text.

numeric literals and array/matrix dimensions in programming code.

numeric literals in code, especially array/matrix dimensions, sizes, and indices within shapes, brackets, and function calls

No Comments

No Known Activations

numbers and code syntax

The neuron detects numeric literals (integer or floating‐point numbers) in code or text.

numeric literals and array/matrix dimensions in programming code.

numeric literals in code, especially array/matrix dimensions, sizes, and indices within shapes, brackets, and function calls

No Comments

No Known Activations