INDEX

Explanations

command line options and targets

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

double hyphen command-line option flags (like `--config`, `--version`, `--help`).

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

Triggered by @bfgcn7frf5

This neuron detects long-form command-line option flags (tokens beginning with “--”).

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

 mortos

-1.99

 detalladas

-1.84

-1.82

 konkrét

-1.63

 motivadoras

-1.62

繻

-1.59

縯

-1.58

 وت

-1.58

 presentó

-1.57

 utilisées

-1.55

POSITIVE LOGITS

マイズ

1.98

 locali

1.80

('--

1.73

理石

1.69

嫲

1.66

suo

1.63

bym

1.60

堉

1.60

的一声

1.57

__',

1.57

Activations Density 0.001%

command line options and targets

double hyphen command-line option flags (like `--config`, `--version`, `--help`).

This neuron detects long-form command-line option flags (tokens beginning with “--”).

No Known Activations

command line options and targets

double hyphen command-line option flags (like `--config`, `--version`, `--help`).

This neuron detects long-form command-line option flags (tokens beginning with “--”).

No Known Activations