INDEX

Explanations

words associated with being singled out or highlighted for distinction

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

hari

-0.14

edback

-0.14

artisan

-0.13

 Fact

-0.13

abeth

-0.13

tein

-0.13

 Address

-0.13

 address

-0.13

ÛĮÙĨÙĩ

-0.13

 Cowboy

-0.13

POSITIVE LOGITS

ä¸ĢåĪĩ

0.14

agram

0.14

á»ĩn

0.14

umba

0.14

GraphNode

0.14

 Tear

0.14

/Area

0.14

.returnValue

0.14

uen

0.14

Activations Density 0.152%

words associated with being singled out or highlighted for distinction

No Comments

No Known Activations