INDEX

Explanations

structured data and identifiers

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

JSON key-value pairs where the key is followed by a colon, particularly in structured data formats.

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

This neuron detects object or map key names (the field names before the colon in JSON‐ or YAML‐style structures).

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

 they

-1.84

-1.57

 camiones

-1.47

 contenedores

-1.46

衚

-1.44

 devrait

-1.42

 alojamientos

-1.41

龶

-1.41

錾

-1.39

回来了

-1.37

POSITIVE LOGITS

of

1.92

 such

1.84

not

1.80

for

1.55

;%%

1.41

any

1.40

endforeach

1.40

now

1.38

 have

1.38

1.35

Activations Density 0.029%

structured data and identifiers

JSON key-value pairs where the key is followed by a colon, particularly in structured data formats.

This neuron detects object or map key names (the field names before the colon in JSON‐ or YAML‐style structures).

No Comments

No Known Activations

structured data and identifiers

JSON key-value pairs where the key is followed by a colon, particularly in structured data formats.

This neuron detects object or map key names (the field names before the colon in JSON‐ or YAML‐style structures).

No Comments

No Known Activations