INDEX

Explanations

instances of the word "where" typically followed by mathematical notation or expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

¤í

-0.15

á»ķi

-0.15

oding

-0.15

AMED

-0.14

azing

-0.14

linkplain

-0.14

 Serif

-0.14

etail

-0.14

loff

-0.14

tro

-0.14

POSITIVE LOGITS

 Downs

0.15

 again

0.14

osi

0.14

 Hamp

0.14

ÐµÑĦ

0.14

 Mint

0.14

aan

0.14

 Tanner

0.13

MOVE

0.13

ì¡±

0.13

Activations Density 0.023%

instances of the word "where" typically followed by mathematical notation or expressions

No Comments

No Known Activations