INDEX

Explanations

specific values or codes related to moderation and user account management

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

vÄĽÅĻ

-0.14

 Fulton

-0.12

Éµ

-0.12

 biomedical

-0.11

 retrospect

-0.11

 å£

-0.11

 Sunshine

-0.11

agento

-0.11

steller

-0.11

eyim

-0.11

POSITIVE LOGITS

 Administration

0.26

 Admin

0.26

 administration

0.25

 ADMIN

0.24

 Server

0.23

 admin

0.23

Administration

0.23

 admins

0.22

 server

0.22

administration

0.21

Activations Density 0.020%

specific values or codes related to moderation and user account management

No Comments

No Known Activations