INDEX

Explanations

blog posts tagged

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

The neuron fires on key content words—especially proper names and other topic-defining nouns—i.e. named entities or main subject terms.

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

punctuation and formatting marks that separate sections or items in lists, particularly commas and line breaks in structured text.

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

blog tag headers/labels, especially the “Posts Tagged” lines indicating tagged topics.

oai_token-act-pair · gpt-5 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

Not in Any Lists

No Comments

Negative Logits

had

-1.34

Other

-1.18

 superbly

-1.16

 Since

-1.13

 undeniably

-1.12

 after

-1.09

 When

-1.08

 gave

-1.07

人に

-1.06

 nestled

-1.06

POSITIVE LOGITS

かっこ

1.40

唖

1.29

かわい

1.25



1.23

 !!!!

1.23

嵓

1.16

だけでなく

1.16

 brilho

1.15

才可以

1.15

カッコ

1.13

Activations Density 0.002%

blog posts tagged

The neuron fires on key content words—especially proper names and other topic-defining nouns—i.e. named entities or main subject terms.

punctuation and formatting marks that separate sections or items in lists, particularly commas and line breaks in structured text.

blog tag headers/labels, especially the “Posts Tagged” lines indicating tagged topics.

No Comments

No Known Activations