INDEX

Explanations

code element names

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

code in statically-typed programming languages, particularly focusing on type declarations and class/namespace definitions.

oai_token-act-pair · claude-4-5-sonnet

Triggered by @bfgcn7frf5

The neuron detects snippets of programming code or code‐like tokens (identifiers, language keywords, API names, types, etc.).

oai_token-act-pair · o4-mini

Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_10/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 сьогодні

-1.16

 сейчас

-1.14

 today

-1.11

 сегодня

-1.11

 довольно

-1.09

 threefold

-1.08

Several

-1.05

now

-1.05

 erschien

-1.04

 three

-1.02

POSITIVE LOGITS

of

1.23

 deltaTime

1.12

 familien

1.12

чами

1.09

 dage

1.09

ënten

1.09

GraphicsUnit

1.08

 ре

1.08

 xPos

1.07

 tille

1.06

Activations Density 0.241%

code element names

code in statically-typed programming languages, particularly focusing on type declarations and class/namespace definitions.

The neuron detects snippets of programming code or code‐like tokens (identifiers, language keywords, API names, types, etc.).

No Known Activations

code element names

code in statically-typed programming languages, particularly focusing on type declarations and class/namespace definitions.

The neuron detects snippets of programming code or code‐like tokens (identifiers, language keywords, API names, types, etc.).

No Known Activations