INDEX

Explanations

punctuation and clauses that indicate conditions or results in programming or logic contexts

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

eld

-0.07

istine

-0.06

mrb

-0.06

esium

-0.06

hta

-0.06

Ø´Ø¯

-0.06

uden

-0.06

ÐµÐ½Ð´

-0.06

ari

-0.06

req

-0.06

POSITIVE LOGITS

 problem

0.09

 when

0.08

 Problem

0.08

 trouble

0.08

ãĤ«ãĥ¼

0.07

 When

0.07

Problem

0.07

ëįĶëĭĪ

0.07

 Trouble

0.07

 nÃ¤r

0.07

Activations Density 0.012%