INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 11-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L11R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.11.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

ri

-0.20

name

-0.20

../../../

-0.16

lation

-0.16

tes

-0.16

rou

-0.16

dest

-0.16

dom

-0.15

iston

-0.15

ric

-0.15

POSITIVE LOGITS

othy

0.22

åĪ»

0.21

arrow

0.19

lessly

0.17

punkt

0.17

ofday

0.17

åĢĻ

0.16

hare

0.16

oris

0.15

ald

0.15

Activations Density 0.206%

No Comments

No Known Activations