INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

aign

-0.16

ario

-0.16

ccc

-0.16

aptop

-0.15

 king

-0.15

Ã¡rio

-0.15

ARTH

-0.14

_secure

-0.14

.Theme

-0.14

ragen

-0.14

POSITIVE LOGITS

 Field

0.19

 field

0.18

 FIELD

0.15

'field

0.15

Field

0.15

ongo

0.15

ichen

0.15

_requested

0.14

endale

0.14

ivery

0.14

Activations Density 0.011%

No Comments

No Known Activations