INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

ucci

-0.15

éĻ

-0.15

 music

-0.15

bine

-0.15

illet

-0.15

ofire

-0.15

uchi

-0.14

agne

-0.14

 Music

-0.13

enso

-0.13

POSITIVE LOGITS

cos

0.19

Cos

0.18

cos

0.18

 yeah

0.17

 bass

0.17

Cos

0.16

 fuck

0.15

 Bass

0.15

 Tony

0.15

 Sabbath

0.15

Activations Density 0.000%

No Known Activations

This feature has no known activations.