INDEX

Explanations

programming keywords and functions

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

underscores in function names, variable names, and identifiers in code.

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

Triggered by @bfgcn7frf5

This neuron detects code identifier prefixes immediately before an underscore in function or method names (e.g. “get_”, “add_”, “infer_”, “write_”).

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

 моей

-1.98

 моих

-1.84

 последних

-1.77

 ligeramente

-1.72

-1.69

 beinahe

-1.67

 некоторых

-1.67

 последние

-1.63

 Posteriormente

-1.63

 последнее

-1.61

POSITIVE LOGITS

 such

2.70

the

2.50

of

2.34

 from

2.33

2.23

 with

2.19

 what

1.94

 something

1.89

in

1.87

 their

1.80

Activations Density 0.011%

programming keywords and functions

underscores in function names, variable names, and identifiers in code.

This neuron detects code identifier prefixes immediately before an underscore in function or method names (e.g. “get_”, “add_”, “infer_”, “write_”).

No Known Activations

programming keywords and functions

underscores in function names, variable names, and identifiers in code.

This neuron detects code identifier prefixes immediately before an underscore in function or method names (e.g. “get_”, “add_”, “infer_”, “write_”).

No Known Activations