INDEX

Explanations

expressions of shame or embarrassment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

im

-0.19

 å¿

-0.15

 Kings

-0.14

imli

-0.14

Î½Ïİ

-0.14

Ã¡le

-0.14

ature

-0.14

 alleg

-0.13

Zub

-0.13

POSITIVE LOGITS

sh

0.29

enan

0.27

rou

0.22

odd

0.20

amed

0.19

abb

0.19

aria

0.19

abby

0.18

rew

0.18

amb

0.18

Activations Density 0.013%

expressions of shame or embarrassment

No Comments

No Known Activations