INDEX

Explanations

good morning, good old, good luck, good times

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

The neuron primarily detects the word “good” (in any capitalization).

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

the word "Good" when it appears as part of a proper noun or brand name.

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

positive evaluative language expressing goodness or quality, especially when it forms part of titles, brand names, or fixed phrases built on that root.

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

 Clique

0.69

σε

0.68

ılım

0.64

 utmost

0.63

 Clive

0.63

ının

0.62

gerald

0.61

 worded

0.61

니다

0.60

 fennel

0.59

POSITIVE LOGITS

GOOD

1.07

好的

1.04

 GOOD

0.96

Good

0.93

 अच्छ

0.92

 good

0.91

 доб

0.91

good

0.89

 goede

0.87

 טוב

0.84

Activations Density 0.128%

good morning, good old, good luck, good times

The neuron primarily detects the word “good” (in any capitalization).

the word "Good" when it appears as part of a proper noun or brand name.

positive evaluative language expressing goodness or quality, especially when it forms part of titles, brand names, or fixed phrases built on that root.

No Comments

No Known Activations

good morning, good old, good luck, good times

The neuron primarily detects the word “good” (in any capitalization).

the word "Good" when it appears as part of a proper noun or brand name.

positive evaluative language expressing goodness or quality, especially when it forms part of titles, brand names, or fixed phrases built on that root.

No Comments

No Known Activations