INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

Ø±Ø²

-0.14

unders

-0.14

uw

-0.14

swire

-0.13

asti

-0.13

æĿĤ

-0.13

 sediment

-0.13

 reuse

-0.13

 unders

-0.13

é»

-0.13

POSITIVE LOGITS

pod

0.43

pod

0.41

 podcast

0.40

 podcasts

0.39

Pod

0.38

 Podcast

0.37

Pod

0.36

_pod

0.33

 pods

0.33

 episode

0.31

Activations Density 0.000%

No Known Activations

This feature has no known activations.