INDEX

Explanations

bracketed structures or mathematical expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 13-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L13R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.13.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

Ø§Ø¨ÛĮ

-0.17

-0.15

sian

-0.14

ersiz

-0.14

sak

-0.13

sah

-0.13

ÑĹ

-0.13

olem

-0.13

eldorf

-0.13

LOBAL

-0.13

POSITIVE LOGITS

Â¯

0.14

chio

0.14

ayette

0.14

 compos

0.14

 SENT

0.14

cape

0.13

 láº¡c

0.13

 berk

0.13

arto

0.13

allo

0.13

Activations Density 0.024%

bracketed structures or mathematical expressions

No Comments

No Known Activations