INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

alim

-0.17

atham

-0.16

ascript

-0.15

leared

-0.14

 crossorigin

-0.14

 ultimately

-0.14

ErrorHandler

-0.14

à¸Ńà¸ģà¸Īà¸²à¸ģà¸Ļ

-0.13

inu

-0.13

oga

-0.12

POSITIVE LOGITS

tplib

0.13

iki

0.13

 Wiki

0.13

”

0.13

ikip

0.13

su

0.12

 Tucker

0.12

Ø¬ÙĪÛĮ

0.12

lev

0.12

Kee

0.12

Activations Density 0.119%

No Comments

No Known Activations