INDEX

Explanations

string theory

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

This neuron spikes on uncommon or low‐frequency content tokens (i.e. rare words or subword pieces).

oai_token-act-pair · o4-mini Triggered by @jyhe0408

the word "not" or negations in technical, instructional, or formal contexts.

oai_token-act-pair · claude-4-5-sonnet Triggered by @jyhe0408

phrases indicating problems, failures, or absence, often in technical or communication contexts.

oai_token-act-pair · gpt-5 Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

י

0.77

Ö

0.70

 neural

0.70

 тех

0.70

PLN

0.70

 ран

0.69

 Ин

0.69

TUT

0.69

 кар

0.68

 ۱۴

0.67

POSITIVE LOGITS

}])

0.75

聞き

0.75

читься

0.74

 deduce

0.74

}`

0.73

消毒

0.70

 কাশ্ম

0.70

érieurs

0.69

差し

0.69

而去

0.69

Activations Density 0.000%

string theory

This neuron spikes on uncommon or low‐frequency content tokens (i.e. rare words or subword pieces).

the word "not" or negations in technical, instructional, or formal contexts.

phrases indicating problems, failures, or absence, often in technical or communication contexts.

No Comments

No Known Activations

string theory

This neuron spikes on uncommon or low‐frequency content tokens (i.e. rare words or subword pieces).

the word "not" or negations in technical, instructional, or formal contexts.

phrases indicating problems, failures, or absence, often in technical or communication contexts.

No Comments

No Known Activations