INDEX

Explanations

The neuron strongly fires on numeric tokens (especially floating‐point or measurement‐style numbers).

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

the word "please" or similar polite imperative phrases in informal, low-quality text.

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

phrases typical of product listings or marketing copy, such as model identifiers, quality descriptors, fit/sizing terms, shipping/transaction language, and polite calls to action.

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

 bumi

0.81

 operaciones

0.80

 expediency

0.79

 badań

0.78

 analizar

0.78

 accéder

0.75

Kč

0.73

 Ingeniería

0.73

 Versicherung

0.73

 anggaran

0.73

POSITIVE LOGITS

ਕ

0.89

야

0.82

囉

0.75

ტ

0.74

と思って

0.73

라고

0.72

ጠቀም

0.71

ਸ

0.70

కీ

0.70

이라면

0.69

Activations Density 0.000%

The neuron strongly fires on numeric tokens (especially floating‐point or measurement‐style numbers).

the word "please" or similar polite imperative phrases in informal, low-quality text.

phrases typical of product listings or marketing copy, such as model identifiers, quality descriptors, fit/sizing terms, shipping/transaction language, and polite calls to action.

No Comments

No Known Activations

The neuron strongly fires on numeric tokens (especially floating‐point or measurement‐style numbers).

the word "please" or similar polite imperative phrases in informal, low-quality text.

phrases typical of product listings or marketing copy, such as model identifiers, quality descriptors, fit/sizing terms, shipping/transaction language, and polite calls to action.

No Comments

No Known Activations