INDEX

Explanations

heterogeneous subtypes and gases

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

words ending in "ing" or plural words ending in "es" or "ans".

oai_token-act-pair · claude-3-7-sonnet-20250219 Triggered by @neilrathi

The neuron strongly activates on subword suffixes at the ends of words (i.e. morphological endings).

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

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_10/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

</b>

-2.17

籏

-2.06

+}$

-1.90

เพราะ

-1.85

铪

-1.80

礻

-1.79

萣

-1.74

 GenerationType

-1.70

ванием

-1.70

杙

-1.69

POSITIVE LOGITS

了你

1.70

 berdua

1.61

Ꮂ

1.55

 benefitted

1.48

 informiert

1.48

but

1.45

There

1.45

 Dritte

1.44

 monumento

1.41

時間で

1.41

Activations Density 0.072%

heterogeneous subtypes and gases

words ending in "ing" or plural words ending in "es" or "ans".

The neuron strongly activates on subword suffixes at the ends of words (i.e. morphological endings).

No Comments

No Known Activations

heterogeneous subtypes and gases

words ending in "ing" or plural words ending in "es" or "ans".

The neuron strongly activates on subword suffixes at the ends of words (i.e. morphological endings).

No Comments

No Known Activations