INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 10-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L10R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.10.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

 dual

-0.14

Laf

-0.14

ield

-0.14

lan

-0.13

 Vern

-0.13

Lan

-0.13

IELD

-0.12

 Trib

-0.12

POSITIVE LOGITS

#ac

0.17

adel

0.17

ffe

0.16

.GroupLayout

0.15

adlo

0.15

eyse

0.15

ereco

0.15

fol

0.15

#aa

0.15

atel

0.14

Activations Density 2.564%

No Comments

No Known Activations