INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

azen

-0.18

localized

-0.16

 DISPATCH

-0.15

 dispatch

-0.15

neighbor

-0.15

 neighbors

-0.15

spread

-0.14

 Neighbor

-0.14

dispatch

-0.14

 neighborhood

-0.14

POSITIVE LOGITS

 professional

0.30

professional

0.27

 Professional

0.26

 professionnel

0.24

Professional

0.24

 professionally

0.23

 professionals

0.22

 professionalism

0.21

 profession

0.20

 Profession

0.20

Activations Density 0.000%

No Known Activations

This feature has no known activations.