INDEX

Explanations

references to news articles and their rankings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

ãĤ»ãĥ³ãĤ¿ãĥ¼

-0.17

æ¾¤

-0.15

umen

-0.15

 Grat

-0.15

arra

-0.14

kem

-0.14

omen

-0.14

Wunused

-0.14

ummer

-0.14

Îĸ

-0.14

POSITIVE LOGITS

 News

0.43

News

0.39

news

0.39

 news

0.39

 NEWS

0.32

-news

0.31

_news

0.30

NEWS

0.27

/news

0.27

.news

0.27

Activations Density 0.030%

references to news articles and their rankings

No Comments

No Known Activations

references to news articles and their rankings

No Comments

No Known Activations