INDEX

Explanations

phrases related to groups and collective identities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 25-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L25R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.25.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

ä¸Ģä¸ªäºº

-0.22

someone

-0.21

 someone

-0.20

 itself

-0.20

çļĦä¸Ģä¸ª

-0.20

æĺ¯ä¸Ģä¸ª

-0.19

 Someone

-0.18

 ÎµÎ½ÏĮÏĤ

-0.17

 jemand

-0.16

ä¸Ģä¸ª

-0.16

POSITIVE LOGITS

 themselves

0.45

 thems

0.22

äºĽ

0.21

 ones

0.21

 yourselves

0.21

 members

0.20

 ÑģÐ°Ð¼Ð¸

0.19

 nhá»¯ng

0.18

 extensions

0.18

 mere

0.18

Activations Density 0.385%

phrases related to groups and collective identities

No Comments

No Known Activations