INDEX

Explanations

probability and estimation

The neuron fires on hedging or speculative words and adverbs—i.e. terms expressing approximation, probability, or opinion (like “probably,” “should,” “roughly,” “denken/думаю,” etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 perna

-1.62

new

-1.58

 gainera

-1.48

 publié

-1.46



-1.43

鉏

-1.41

傒

-1.41

喁

-1.40

𓆏

-1.38

-1.36

POSITIVE LOGITS

of

1.45

 stasiun

1.45

 apsau

1.45

 with

1.44

ökning

1.41

 atlik

1.40

 einzigen

1.38

殒

1.35

 galima

1.34

↵↵↵↵↵↵↵↵↵↵

1.34

Activations Density 0.004%