INDEX

Explanations

references to Pokémon training and competitions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

XB

-0.17

xb

-0.16

jec

-0.15

Men

-0.15

 patri

-0.14

men

-0.14

ë¡Ģ

-0.14

æĿĢ

-0.14

 elast

-0.14

pig

-0.14

POSITIVE LOGITS

 Trainer

0.24

 PokÃ©mon

0.23

trainer

0.22

Gym

0.21

 Psychic

0.21

 Pokemon

0.20

 trainers

0.20

gym

0.19

Ã©mon

0.19

 trainer

0.19

Activations Density 0.036%

references to Pokémon training and competitions

No Comments

No Known Activations