INDEX

Explanations

terms related to performance metrics and evaluations

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

 ì¹ĺ

-0.15

usercontent

-0.15

 exempt

-0.15

 cáº¥u

-0.14

çĳŁ

-0.14

Ø§ÙĤÙĦ

-0.14

endas

-0.14

æ°ı

-0.14

pson

-0.13

reta

-0.13

POSITIVE LOGITS

 improvement

0.27

 improvements

0.24

/security

0.23

 Improvement

0.22

/us

0.22

 trade

0.20

 gains

0.20

 issues

0.19

 gain

0.18

enh

0.18

Activations Density 0.117%

terms related to performance metrics and evaluations

No Comments

No Known Activations