INDEX

Explanations

concepts, categories, or descriptions

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

This neuron responds to uncommon, content-bearing words (i.e. long or rare topic-specific terms with high information content).

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

statements in instructional or promotional text that use formal transitions and colon-headed structure to introduce explanations, conclusions, or feature summaries.

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

Not in Any Lists

No Comments

Negative Logits

iliśmy

0.68

�

0.66

0.64

ayama

0.64

WOW

0.64

 बदमाशों

0.64

”

0.64

0.63

isman

0.62

POSITIVE LOGITS

 gestão

0.72

 peculiarly

0.66

 crux

0.60

🛠

0.59

❙

0.59

initState

0.57

 notably

0.57

 præ

0.57

वारीत

0.57

很是

0.57

Activations Density 0.001%

concepts, categories, or descriptions

This neuron responds to uncommon, content-bearing words (i.e. long or rare topic-specific terms with high information content).

statements in instructional or promotional text that use formal transitions and colon-headed structure to introduce explanations, conclusions, or feature summaries.

No Comments

No Known Activations

concepts, categories, or descriptions

This neuron responds to uncommon, content-bearing words (i.e. long or rare topic-specific terms with high information content).

statements in instructional or promotional text that use formal transitions and colon-headed structure to introduce explanations, conclusions, or feature summaries.

No Comments

No Known Activations