INDEX

Explanations

subscriptionId, Name, undefined, null, Error

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

attribute or property names in code and structured data formats.

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

This neuron activates on code‐style identifier tokens—particularly names of AST node types and other programming identifiers (often containing underscores or CamelCase).

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

its

-1.77

𓇼

-1.53

ísimas

-1.36

卧槽

-1.34

Its

-1.34

 में

-1.34

絰

-1.34

licitations

-1.30

anything

-1.29

 producteur

-1.28

POSITIVE LOGITS

AppCompatTheme

1.44

 evidentemente

1.36

we

1.34

ardı

1.33

 però

1.27

或者

1.24

 nossos

1.21

 different

1.21

 podríamos

1.18

 différents

1.17

Activations Density 0.002%

subscriptionId, Name, undefined, null, Error

attribute or property names in code and structured data formats.

This neuron activates on code‐style identifier tokens—particularly names of AST node types and other programming identifiers (often containing underscores or CamelCase).

No Comments

No Known Activations

subscriptionId, Name, undefined, null, Error

attribute or property names in code and structured data formats.

This neuron activates on code‐style identifier tokens—particularly names of AST node types and other programming identifiers (often containing underscores or CamelCase).

No Comments

No Known Activations