INDEX

Explanations

modes

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

The neuron is specialized in spotting numeric tokens—numbers, dates, and decimal‐formatted values.

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

very large numbers (typically in the thousands).

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

calls-to-action and link-like anchor text in web-style content, including clickable prompts and event/date references commonly used for navigation.

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

New Auto-Interp

Configuration

google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ప్రభుత్వం

0.78

 rapporto

0.77

పాల

0.77

ผู้

0.77

 confertim

0.77

 पेशेवरों

0.76

 peroneal

0.76

లిన

0.76

および

0.75

 counterv

0.75

POSITIVE LOGITS

뎅

0.69

 putt

0.67

 fetch

0.67

DJ

0.66

 اه

0.66

 Гри

0.66

 اب

0.64

DT

0.64

 тут

0.64

 جنگ

0.64

Activations Density 0.000%

modes

The neuron is specialized in spotting numeric tokens—numbers, dates, and decimal‐formatted values.

very large numbers (typically in the thousands).

calls-to-action and link-like anchor text in web-style content, including clickable prompts and event/date references commonly used for navigation.

No Comments

No Known Activations

modes

The neuron is specialized in spotting numeric tokens—numbers, dates, and decimal‐formatted values.

very large numbers (typically in the thousands).

calls-to-action and link-like anchor text in web-style content, including clickable prompts and event/date references commonly used for navigation.

No Comments

No Known Activations