INDEX

Explanations

code abbreviations

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

language-specific namespaces or package identifiers in code (like abbreviations or short identifiers used in import statements, namespace declarations, or class paths).

oai_token-act-pair · claude-4-5-sonnet Triggered by @bfgcn7frf5

The neuron strongly activates on standalone code‐identifier tokens—especially uppercase acronyms or proper module/namespace names (e.g. ATC, KlayGE, zh, elle).

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

 occhi

-1.46

viewWillAppear

-1.41

lampa

-1.41

 orari

-1.39

Cześć

-1.36

 vicina

-1.36

 dicendo

-1.34

 dicono

-1.34

 industriale

-1.34

簡

-1.33

POSITIVE LOGITS

sistence

1.58

了一眼

1.51

–

1.45

("

1.43

1.41

1.36

1.32

see

1.31

：……

1.30

where

1.28

Activations Density 0.086%

code abbreviations

language-specific namespaces or package identifiers in code (like abbreviations or short identifiers used in import statements, namespace declarations, or class paths).

The neuron strongly activates on standalone code‐identifier tokens—especially uppercase acronyms or proper module/namespace names (e.g. ATC, KlayGE, zh, elle).

No Comments

No Known Activations

code abbreviations

language-specific namespaces or package identifiers in code (like abbreviations or short identifiers used in import statements, namespace declarations, or class paths).

The neuron strongly activates on standalone code‐identifier tokens—especially uppercase acronyms or proper module/namespace names (e.g. ATC, KlayGE, zh, elle).

No Comments

No Known Activations