INDEX

Explanations

references to the concept of "special" or unique attributes and experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

 special

-0.22

 especial

-0.20

hra

-0.19

soever

-0.19

 specific

-0.17

special

-0.16

unes

-0.16

 specials

-0.16

pecific

-0.16

 Special

-0.16

POSITIVE LOGITS

-purpose

0.33

ity

0.30

ties

0.28

ities

0.27

ty

0.26

ised

0.25

izations

0.23

-effects

0.23

ise

0.23

zed

0.23

Activations Density 0.041%

references to the concept of "special" or unique attributes and experiences

No Comments

No Known Activations