INDEX

Explanations

square root and cube root

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

The neuron strongly activates on the markers of numbered or titled sections—e.g. list‐item numerals (“2.”, “7.”, “8.”), step headings, and other sectiontitle tokens—i.e. it detects the beginnings of instruction steps or numbered headings.

oai_token-act-pair · o4-mini 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

 underground

0.64

0.63

 contributing

0.57

Force

0.57

 इंट्रा

0.55

 подзем

0.55

 hart

0.55

 Underground

0.54

ள

0.54

 jabs

0.54

POSITIVE LOGITS

೭

0.62

 الشمس

0.60

 Trigon

0.58

 Rabin

0.58

 Rubi

0.57

ळक

0.56

 coseno

0.56

膳

0.56

地板

0.55

 Fourteenth

0.55

Activations Density 0.000%

square root and cube root

The neuron strongly activates on the markers of numbered or titled sections—e.g. list‐item numerals (“2.”, “7.”, “8.”), step headings, and other sectiontitle tokens—i.e. it detects the beginnings of instruction steps or numbered headings.

No Comments

No Known Activations

square root and cube root

The neuron strongly activates on the markers of numbered or titled sections—e.g. list‐item numerals (“2.”, “7.”, “8.”), step headings, and other sectiontitle tokens—i.e. it detects the beginnings of instruction steps or numbered headings.

No Comments

No Known Activations

square root and cube root

The neuron strongly activates on the markers of numbered or titled sections—e.g. list‐item numerals (“2.”, “7.”, “8.”), step headings, and other section­title tokens—i.e. it detects the beginnings of instruction steps or numbered headings.

No Comments

No Known Activations

square root and cube root

The neuron strongly activates on the markers of numbered or titled sections—e.g. list‐item numerals (“2.”, “7.”, “8.”), step headings, and other section­title tokens—i.e. it detects the beginnings of instruction steps or numbered headings.

No Comments

No Known Activations

The neuron strongly activates on the markers of numbered or titled sections—e.g. list‐item numerals (“2.”, “7.”, “8.”), step headings, and other sectiontitle tokens—i.e. it detects the beginnings of instruction steps or numbered headings.

The neuron strongly activates on the markers of numbered or titled sections—e.g. list‐item numerals (“2.”, “7.”, “8.”), step headings, and other sectiontitle tokens—i.e. it detects the beginnings of instruction steps or numbered headings.