INDEX

Explanations

numeric values and counts that indicate quantities or statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

and

-0.18

and

-0.16

ankind

-0.16

and

-0.16

ovalo

-0.15

ovaly

-0.14

][:

-0.14

ä»¥åıĬ

-0.14

ekk

-0.14

vÄĽd

-0.13

POSITIVE LOGITS

 others

0.34

 them

0.28

 ones

0.26

them

0.26

others

0.25

 thereof

0.24

Others

0.23

 Others

0.23

 whose

0.22

 being

0.21

Activations Density 0.148%

numeric values and counts that indicate quantities or statistics

No Comments

No Known Activations