INDEX

Explanations

terms related to physical training and performance

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

 Ange

-0.06

ypo

-0.06

ys

-0.06

avian

-0.06

 educated

-0.06

gang

-0.06

 Mate

-0.05

 experiment

-0.05

 Falcon

-0.05

 invention

-0.05

POSITIVE LOGITS

 training

0.12

-training

0.11

 Training

0.10

training

0.10

Training

0.09

 trainers

0.09

 TRAIN

0.08

/Dk

0.08

plevel

0.08

 train

0.08

Activations Density 0.005%