INDEX

Explanations

nations, roles, or entities

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

This neuron primarily detects numeric or statistical tokens (e.g. multidigit numbers, scores, years, percentages) in the text.

oai_token-act-pair · o4-mini

Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_22/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 Prenez

-1.00

:{

-0.98

 vimos

-0.91

 Espanha

-0.90

its

-0.88

jekk

-0.87

 then

-0.86

Ἱ

-0.86

 firmas

-0.86

 Kapelle

-0.86

POSITIVE LOGITS

 jūsų

1.00

 osserva

0.95

who

0.95

 mūsų

0.94

ாள

0.93

 aang

0.92

 derniers

0.92

gerufen

0.91

uch

0.91

也是

0.91

Activations Density 0.018%

nations, roles, or entities

This neuron primarily detects numeric or statistical tokens (e.g. multidigit numbers, scores, years, percentages) in the text.

No Known Activations

nations, roles, or entities

This neuron primarily detects numeric or statistical tokens (e.g. multidigit numbers, scores, years, percentages) in the text.

No Known Activations

nations, roles, or entities

This neuron primarily detects numeric or statistical tokens (e.g. multi­digit numbers, scores, years, percentages) in the text.

No Known Activations

nations, roles, or entities

This neuron primarily detects numeric or statistical tokens (e.g. multi­digit numbers, scores, years, percentages) in the text.

No Known Activations

This neuron primarily detects numeric or statistical tokens (e.g. multidigit numbers, scores, years, percentages) in the text.

This neuron primarily detects numeric or statistical tokens (e.g. multidigit numbers, scores, years, percentages) in the text.