INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

anes

-0.20

oron

-0.14

bara

-0.14

veau

-0.14

agli

-0.14

ledged

-0.13

enza

-0.13

eko

-0.13

actory

-0.13

mour

-0.13

POSITIVE LOGITS

adele

0.17

.pix

0.16

errupted

0.14

lights

0.14

sembler

0.13

enticated

0.13

vil

0.13

 forn

0.13

oved

0.13

vik

0.13

Activations Density 0.025%

No Comments

No Known Activations