INDEX

Explanations

instances of the letter "J" or its variants in different contexts

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

duct

-0.27

ductor

-0.25

ductive

-0.22

der

-0.21

verte

-0.21

cul

-0.20

verting

-0.19

compat

-0.19

clusions

-0.18

ferences

-0.17

POSITIVE LOGITS

ÑħÐ¾Ð´

0.26

ÑģÐ»ÐµÐ´

0.24

ÑģÑĤÑĥÐ¿

0.24

Ð¿Ð¾Ñģ

0.23

her

0.22

ÑģÐºÐ¾ÑĢ

0.22

Å¡e

0.21

ÑĢÐµÐ¼ÐµÐ½Ð½Ð¾

0.21

hib

0.21

ability

0.21

Activations Density 0.008%

instances of the letter "J" or its variants in different contexts

No Comments

No Known Activations