INDEX

Explanations

The neuron is keyed to sentences from Chinese local news reports, especially government or police notifications and investigations.

oai_token-act-pair · o4-mini

Triggered by @jyhe0408

people reporting scientific results

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

formal or academic text, particularly sections of research papers, reports, and technical documents.

oai_token-act-pair · claude-3-7-sonnet-20250219

Triggered by @neilrathi

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

ׇ

-1.27

 свої

-1.23

حياته

-1.20

 probably

-1.17

了什么

-1.15

 město

-1.14

を訪

-1.13

 perhaps

-1.13

Ghz

-1.09

 certainly

-1.05

POSITIVE LOGITS

ִּ

1.24

 PLAINT

1.20

 principalement

1.20

וּ

1.16

 ISLANDS

1.16

ְּ

1.15

 APPELL

1.15

人が

1.12

 supprimer

1.11

ָּ

1.09

Activations Density 0.030%

The neuron is keyed to sentences from Chinese local news reports, especially government or police notifications and investigations.

people reporting scientific results

formal or academic text, particularly sections of research papers, reports, and technical documents.

No Known Activations

The neuron is keyed to sentences from Chinese local news reports, especially government or police notifications and investigations.

people reporting scientific results

formal or academic text, particularly sections of research papers, reports, and technical documents.

No Known Activations