INDEX

Explanations

detailed instructions and tutorials for completing tasks

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Negative Logits

rap

-0.06

 Assertion

-0.06

umper

-0.06

ï¼ł

-0.06

aware

-0.06

eniable

-0.06

ffff

-0.05

 Kern

-0.05

urve

-0.05

rve

-0.05

POSITIVE LOGITS

 instruction

0.10

 step

0.10

 instructions

0.10

 instruct

0.09

-step

0.08

 steps

0.08

ä¸ĢæŃ¥

0.08

step

0.07

 Step

0.07

Instruction

0.07

Activations Density 0.022%

No Known Activations