INDEX

Explanations

phrases that indicate composition or assembly

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

Negative Logits

isay

-0.15

Ð¸ÑģÑģ

-0.15

exampleInput

-0.15

UTIL

-0.15

---</

-0.15

.appspot

-0.14

abaj

-0.14

cko

-0.14

ALES

-0.14

tsky

-0.14

POSITIVE LOGITS

leine

0.37

ira

0.26

 possible

0.24

 worse

0.21

IRA

0.21

up

0.20

-for

0.19

 posible

0.19

 sense

0.18

 abund

0.18

Activations Density 0.028%

phrases that indicate composition or assembly

No Known Activations