INDEX

Explanations

exiting loops and conditions

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

the keyword "break" in programming code, particularly at the end of loop or control flow statements.

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

The neuron is highly activated by loop‐control keywords—especially tokens that break out of or exit loops (e.g. “break” and “exit”).

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

 oliveira

-1.57

eſ

-1.50

 preocupado

-1.47

薅

-1.47

⬙

-1.45

⢅

-1.45

 acontec

-1.44

σα

-1.40

 abstrato

-1.36

Lindo

-1.36

POSITIVE LOGITS

1.71

1.63

</b>

1.59

 also

1.55

are

1.51

</u>

1.50

1.41

})$.

1.40

To

1.39

Type

1.39

Activations Density 0.005%

exiting loops and conditions

the keyword "break" in programming code, particularly at the end of loop or control flow statements.

The neuron is highly activated by loop‐control keywords—especially tokens that break out of or exit loops (e.g. “break” and “exit”).

No Comments

No Known Activations

exiting loops and conditions

the keyword "break" in programming code, particularly at the end of loop or control flow statements.

The neuron is highly activated by loop‐control keywords—especially tokens that break out of or exit loops (e.g. “break” and “exit”).

No Comments

No Known Activations