INDEX

Explanations

unique or specialized terms related to specific subjects or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

mÃ¡n

-0.16

Ã©nÃ©

-0.16

Ð¸ÑĩÐ½Ð°

-0.15

Ð»ÐµÐ½Ñĭ

-0.15

AMED

-0.14

andas

-0.14

 Wander

-0.14

herits

-0.14

itto

-0.14

geois

-0.14

POSITIVE LOGITS

izing

0.30

ÃŃcÃŃ

0.28

sing

0.28

Ð½Ð¾ÑģÑı

0.28

ulating

0.27

-ing

0.27

ying

0.27

ising

0.26

ding

0.26

ÂŃing

0.26

Activations Density 0.021%

unique or specialized terms related to specific subjects or concepts

No Comments

No Known Activations