INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

pij

-0.19

ä¸´

-0.15

hoo

-0.15

Ð°Ð»Ð¾Ð²

-0.14

å±ĭ

-0.14

èĩ¨

-0.14

cisi

-0.14

.Nav

-0.13

REFIX

-0.13

azzi

-0.13

POSITIVE LOGITS

 lawn

0.27

 Lawn

0.27

 chains

0.27

 hedge

0.26

 mower

0.25

 Chains

0.24

 Weed

0.23

 trim

0.22

law

0.22

 Toro

0.21

Activations Density 0.022%

No Comments

No Known Activations