INDEX

Explanations

recurring concepts and comparisons in arguments related to education and performance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

 itself

-0.16

 besonders

-0.15

amac

-0.15

ilogy

-0.15

 alone

-0.15

ilib

-0.14

 beyond

-0.14

ãģŁãĤĬ

-0.14

Ã¸j

-0.14

eyond

-0.14

POSITIVE LOGITS

 except

0.27

 than

0.26

minus

0.23

except

0.23

 minus

0.22

 identical

0.20

Except

0.20

 Except

0.20

same

0.19

than

0.18

Activations Density 0.091%

recurring concepts and comparisons in arguments related to education and performance

No Comments

No Known Activations