INDEX

Explanations

words and phrases that indicate caution or carefulness

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

herits

-0.18

aginator

-0.16

Ð¾Ð±ÑĢÐ°Ð·

-0.15

Î¼Î¿Ïģ

-0.15

imal

-0.15

ÑĢÐµÑĪ

-0.14

IMAL

-0.14

oeff

-0.14

á»©

-0.14

clair

-0.14

POSITIVE LOGITS

lest

0.27

ness

0.26

NESS

0.22

/select

0.20

 consideration

0.20

 attention

0.17

íŀĪ

0.17

 avoiding

0.17

ç¿¼

0.17

nes

0.16

Activations Density 0.031%

words and phrases that indicate caution or carefulness

No Comments

No Known Activations