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

ilir

-0.15

Ø§ÛĮÛĮ

-0.14

asca

-0.14

assin

-0.14

erman

-0.13

rary

-0.13

asal

-0.13

ãĥ¡ãĥ©

-0.13

kod

-0.13

odon

-0.13

POSITIVE LOGITS

 narr

0.18

berra

0.15

 zoom

0.14

 stip

0.14

folk

0.14

zoom

0.14

 hypoc

0.14

ifes

0.14

 legisl

0.14

 Ð¿ÑĢÐ¸ÑĤ

0.13

Activations Density 0.000%

No Known Activations

This feature has no known activations.