INDEX

Explanations

Python code imports

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

The neuron fires on occurrences of “double‐underscore” names or tokens (e.g. name, init, FILE) in code.

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

Negative Logits

-2.53

-2.36

 mirando

-2.27

 poderoso

-2.20

-2.16

and

-2.09

 imidlertid

-2.08

 夏季

-2.06

 preparando

-2.06

 चाहता

-2.03

POSITIVE LOGITS

窳

3.13

 vollen

2.36

</u>

2.36

rakech

2.31

 These

2.30

水が

2.25

瓩

2.25

这些人

2.25

 машине

2.23

珵

2.23

Activations Density 0.002%

Python code imports

The neuron fires on occurrences of “double‐underscore” names or tokens (e.g. name, init, FILE) in code.

No Known Activations

Python code imports

The neuron fires on occurrences of “double‐underscore” names or tokens (e.g. name, init, FILE) in code.

No Known Activations

Python code imports

The neuron fires on occurrences of “double‐underscore” names or tokens (e.g. __name__, __init__, __FILE__) in code.

No Known Activations

Python code imports

The neuron fires on occurrences of “double‐underscore” names or tokens (e.g. __name__, __init__, __FILE__) in code.

No Known Activations

The neuron fires on occurrences of “double‐underscore” names or tokens (e.g. name, init, FILE) in code.

The neuron fires on occurrences of “double‐underscore” names or tokens (e.g. name, init, FILE) in code.