INDEX

Explanations

references to titles or figures associated with authority and leadership

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

bine

-0.18

hread

-0.18

è

-0.17

à¸ģà¸¥à¸²à¸ĩ

-0.15

eros

-0.15

ullah

-0.15

eydi

-0.15

å´

-0.15

uropean

-0.15

oose

-0.14

POSITIVE LOGITS

ship

0.36

ships

0.34

kip

0.23

lings

0.22

SHIP

0.21

 Protector

0.21

 Mayor

0.20

 Voldemort

0.20

 Lieutenant

0.20

ly

0.19

Activations Density 0.015%

references to titles or figures associated with authority and leadership

No Comments

No Known Activations