INDEX

Explanations

The neuron fires on strong, goal-oriented action verbs and terms used in planning or prioritizing tasks (e.g., identify, clarify, set, plan, start, prioritize, help, understand).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 उसको

1.17

 उसने

1.17

 उसे

1.14

 내가

1.05

 इसने

1.05

 ಅವನು

1.04

 얘가

1.04

 ওকে

1.03

 nisam

1.01

 ему

1.00

POSITIVE LOGITS

their

4.69

 their

4.63

 leurs

4.38

Their

4.38

他们的

4.28

他們的

4.22

 их

4.06

 themselves

4.00

 Their

3.98

 deres

3.98

Activations Density 4.253%