INDEX

Explanations

listing entries or items

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

The neuron strongly activates on tokens that express quantities or numbers—digits and words indicating amounts (e.g. “2,” “loads,” “all,” “some,” “two,” “number,” etc.).

oai_token-act-pair · o4-mini

Triggered by @jyhe0408

determiners and quantifiers preceding nouns in descriptive or instructional contexts.

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

Triggered by @jyhe0408

expressions that indicate quantities or enumeration—such as selecting, segmenting, or listing items (e.g., “all,” “some,” “each,” “one/two,” “example,” “any,” “different,” “separate”).

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

Negative Logits

Components

0.68

 Components

0.65

 δη

0.57

 COMPONENTS

0.55

 corticosteroids

0.55

 لوا

0.55

 composants

0.54

 pharmac

0.54

 аксе

0.54

FormParams

0.54

POSITIVE LOGITS

 entries

0.95

 작품

0.92

 storie

0.89

 songs

0.87

 handful

0.87

 submissions

0.82

作品

0.82

 stories

0.81

案件

0.80

 एखा

0.79

Activations Density 0.750%

listing entries or items

The neuron strongly activates on tokens that express quantities or numbers—digits and words indicating amounts (e.g. “2,” “loads,” “all,” “some,” “two,” “number,” etc.).

determiners and quantifiers preceding nouns in descriptive or instructional contexts.

expressions that indicate quantities or enumeration—such as selecting, segmenting, or listing items (e.g., “all,” “some,” “each,” “one/two,” “example,” “any,” “different,” “separate”).

No Known Activations