INDEX

Explanations

elements related to design and geometry

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

raq

-0.15

sak

-0.14

ìļ©

-0.14

ÙİØ³

-0.13

 cubes

-0.13

kke

-0.13

Ã¤m

-0.13

 rects

-0.13

Duy

-0.13

eldorf

-0.13

POSITIVE LOGITS

 shape

0.59

-shaped

0.50

shape

0.47

 Shape

0.42

 shaped

0.40

 design

0.40

 shapes

0.39

å½¢

0.39

Shape

0.38

_shape

0.38

Activations Density 0.303%

elements related to design and geometry

No Comments

No Known Activations