INDEX

Explanations

terms that express a sense of being exceptional or notable

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 20-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L20R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.20.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

dek

-0.14

aved

-0.14

ãģŁãĤģãģ®

-0.14

acier

-0.14

bie

-0.14

ster

-0.14

essa

-0.14

 major

-0.14

stery

-0.13

ney

-0.13

POSITIVE LOGITS

 amount

0.27

 amounts

0.26

ly

0.24

LY

0.23

 Amount

0.21

amount

0.21

Amount

0.18

íŀĪ

0.18

ingly

0.18

mente

0.18

Activations Density 0.083%

terms that express a sense of being exceptional or notable

No Comments

No Known Activations