INDEX

Explanations

This neuron detects direct reader address—tokens that refer to or speak to “you,” “anyone,” or “someone.”

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

phrases directly addressing or targeting the reader, typically in promotional or instructional contexts.

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

phrases that directly address or define a target audience in promotional or informational copy (e.g., second-person “you” and broad audience descriptors).

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

New Auto-Interp

Configuration

google/gemma-scope-2-12b-pt/resid_post/layer_31_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

 গাঙ্গুল

0.58

 testes

0.58

 inputted

0.57

 훨씬

0.56

 usamos

0.55

 लाश

0.55

spersed

0.55

 acceptable

0.54

ктак

0.53

ritus

0.53

POSITIVE LOGITS

 wanting

2.02

 interested

1.90

 seeking

1.84

 muốn

1.82

 aspiring

1.78

想要

1.77

Interested

1.72

 ingin

1.70

interested

1.70

 desirous

1.70

Activations Density 1.023%

This neuron detects direct reader address—tokens that refer to or speak to “you,” “anyone,” or “someone.”

phrases directly addressing or targeting the reader, typically in promotional or instructional contexts.

phrases that directly address or define a target audience in promotional or informational copy (e.g., second-person “you” and broad audience descriptors).

No Comments

No Known Activations

This neuron detects direct reader address—tokens that refer to or speak to “you,” “anyone,” or “someone.”

phrases directly addressing or targeting the reader, typically in promotional or instructional contexts.

phrases that directly address or define a target audience in promotional or informational copy (e.g., second-person “you” and broad audience descriptors).

No Comments

No Known Activations