INDEX

Explanations

code, technical terms, and headers

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

opening quotation marks or quote-like characters in code and technical documentation.

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

The neuron activates on code‐style tokens—particularly uppercase acronyms or identifiers (often containing underscores) typical of programming or technical headers.

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

 those

-1.46

our

-1.43

啦

-1.42

 both

-1.39

 美丽

-1.36

："

-1.34

 leggen

-1.30

erweise

-1.30

 найбіль

-1.29

 wollt

-1.28

POSITIVE LOGITS

 esigen

1.40

 nawet

1.31

на

1.30

1.29

是怎么回事

1.28

 hatta

1.25

</b>

1.24

<h6>

1.20

将来

1.20

Activations Density 0.008%

code, technical terms, and headers

opening quotation marks or quote-like characters in code and technical documentation.

The neuron activates on code‐style tokens—particularly uppercase acronyms or identifiers (often containing underscores) typical of programming or technical headers.

No Comments

No Known Activations

code, technical terms, and headers

opening quotation marks or quote-like characters in code and technical documentation.

The neuron activates on code‐style tokens—particularly uppercase acronyms or identifiers (often containing underscores) typical of programming or technical headers.

No Comments

No Known Activations