INDEX

Explanations

message IDs and references

This neuron detects occurrences of the term “message” (including its abbreviation “msg”), especially in contexts of message fields or identifiers.

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

},

-1.31

xt

-1.20

 خلال

-1.20

 diseñados

-1.16

for

-1.15

↵↵

-1.15

Bà

-1.14

 seepage

-1.12

-1.07

isan

-1.07

POSITIVE LOGITS

 your

1.53

 about

1.45

 isolada

1.23

饰

1.22

 recomendada

1.20

 estrés

1.18

 them

1.16

any

1.16

صاص

1.15

 recent

1.12

Activations Density 0.052%