INDEX

Explanations

change to authentication method

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

The neuron strongly activates on numeric tokens—i.e. standalone numbers or digits (quantities, years, counts).

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

low-quality, spammy, or machine-generated text.

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

metadata- or summary-style tokens in expository text (bibliographic/credit markers, abstract nouns, possessives, years, and similar structural cues).

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

New Auto-Interp

Configuration

google/gemma-scope-2-12b-pt/resid_post/layer_24_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

EST

0.72

png

0.66

tive

0.66

 dúvida

0.64

Ine

0.61

eil

0.61

الأ

0.60

eiro

0.59

flowing

0.59

PATTERN

0.59

POSITIVE LOGITS

扩展

0.83

景点

0.81

损伤

0.80

טי

0.79

凵

0.79

 effectué

0.77

 savor

0.77

车辆

0.76

 végétaux

0.76

去掉

0.76

Activations Density 0.000%

change to authentication method

The neuron strongly activates on numeric tokens—i.e. standalone numbers or digits (quantities, years, counts).

low-quality, spammy, or machine-generated text.

metadata- or summary-style tokens in expository text (bibliographic/credit markers, abstract nouns, possessives, years, and similar structural cues).

No Comments

No Known Activations