INDEX

Explanations

numbers and counting

The neuron fires on numeric tokens—i.e. numbers and numeric expressions in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.63

.//

0.60

,»

0.57

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.56

 yakni

0.56

 بەر

0.56

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.56

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.56

 (…)

0.55

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.55

POSITIVE LOGITS

---

3.36

---

2.72

 ----

1.75

 -----

1.66

----

1.60

-----

1.48

<0xC2>

1.48

---|

1.48

------

1.41

 -------

1.38

Activations Density 1.212%