INDEX

Explanations

quantitative adjectives indicating amounts or quantities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

ynes

-0.14

FFE

-0.14

Ð´ÐµÐ¼

-0.14

ermann

-0.13

aterno

-0.13

maal

-0.13

sett

-0.13

Ð¾Ð³Ð¾Ð´

-0.13

most

-0.13

SharedPtr

-0.13

POSITIVE LOGITS

 respects

0.38

 ways

0.31

 respect

0.30

 sense

0.30

way

0.27

 regards

0.25

 senses

0.24

 cases

0.24

 aspects

0.24

ways

0.22

Activations Density 0.051%

quantitative adjectives indicating amounts or quantities

No Comments

No Known Activations