INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

LDS

-0.15

Gaz

-0.14

BDS

-0.14

 Gibraltar

-0.14

 Styles

-0.13

 Sold

-0.13

 idol

-0.13

 Polit

-0.13

LBL

-0.13

 Seller

-0.13

POSITIVE LOGITS

 Carnegie

0.40

 museum

0.37

 Museum

0.35

 Hill

0.33

 museums

0.32

 Mellon

0.30

 Muse

0.24

useum

0.23

 muse

0.23

 hill

0.23

Activations Density 0.002%

No Comments

No Known Activations