INDEX

Explanations

commands and phrases suggesting vigilance or attention

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

-0.16

atti

-0.15

 âĹĦ

-0.15

ickle

-0.15

 èµ¤

-0.15

oucÃŃ

-0.14

 pace

-0.14

æĿ¿

-0.14

deniz

-0.14

rok

-0.14

POSITIVE LOGITS

 tuned

0.31

Tun

0.28

 tuning

0.24

tun

0.21

 tunes

0.20

 tune

0.19

Cal

0.19

 Tune

0.18

 checking

0.18

cal

0.16

Activations Density 0.023%

commands and phrases suggesting vigilance or attention

No Comments

No Known Activations