INDEX

Explanations

numerical values and references to game mechanics or improvements

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

Ã¼le

-0.15

dispatcher

-0.15

ç°

-0.14

 Thanh

-0.14

 experiment

-0.14

 íĻĺ

-0.14

-eslint

-0.14

rud

-0.13

 reception

-0.13

ĴĮ

-0.13

POSITIVE LOGITS

 cards

0.27

 card

0.24

HQ

0.24

 Cards

0.22

 deck

0.21

 agendas

0.20

agenda

0.20

HQ

0.20

 face

0.20

-card

0.19

Activations Density 0.003%

numerical values and references to game mechanics or improvements

No Comments

No Known Activations