INDEX

Explanations

terms associated with virtue and moral excellence

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

itz

-0.16

throp

-0.15

á»Ĩ

-0.15

atters

-0.15

lds

-0.14

æģ¯

-0.14

arse

-0.14

 xÃ©t

-0.14

hammer

-0.14

erus

-0.14

POSITIVE LOGITS

ually

0.34

uous

0.26

ue

0.24

ues

0.21

tual

0.20

uos

0.19

usize

0.19

uale

0.19

 virt

0.19

uele

0.18

Activations Density 0.003%

terms associated with virtue and moral excellence

No Comments

No Known Activations