INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 23-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L23R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.23.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

Ã³m

-0.16

 conoc

-0.15

izaciÃ³n

-0.15

 Primer

-0.15

isis

-0.14

acciÃ³n

-0.14

 posible

-0.14

Tabla

-0.14

 mientras

-0.14

 desarroll

-0.14

POSITIVE LOGITS

inha

0.21

 Brazilian

0.21

Ã´n

0.20

Ã¡rio

0.20

Ã´

0.19

agem

0.19

Ãªn

0.19

iÃ£o

0.19

iÃ§Ã£o

0.18

 SÃ£o

0.18

Activations Density 0.000%

No Known Activations

This feature has no known activations.