INDEX

Explanations

occurrences of the letter "j" and its variations 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 @ 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

-0.24

sik

-0.18

wt

-0.18

ska

-0.17

sar

-0.17

à¤¨

-0.17

obl

-0.17

sled

-0.16

iÃ³

-0.16

POSITIVE LOGITS

ournals

0.29

0.27

oints

0.26

à¤¼

0.23

ourn

0.23

à¥įà¤ŀ

0.22

upyter

0.22

argon

0.21

OURNAL

0.20

ardin

0.20

Activations Density 0.117%

occurrences of the letter "j" and its variations in different contexts

No Known Activations