INDEX

Explanations

references to educational courses or classes

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

opoulos

-0.20

soever

-0.19

 courses

-0.18

 coursework

-0.18

Courses

-0.18

 Courses

-0.18

 cursos

-0.18

omik

-0.17

_courses

-0.17

lep

-0.16

POSITIVE LOGITS

ware

0.41

wares

0.27

work

0.24

Ware

0.23

book

0.23

mates

0.23

WARE

0.23

books

0.21

mate

0.20

 ware

0.20

Activations Density 0.032%

references to educational courses or classes

No Comments

No Known Activations