INDEX

Explanations

mentions of the word "Ju" and its variations, likely signaling references to juvenile or youth-related 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

æĿ

-0.17

ycz

-0.17

ijk

-0.17

rokes

-0.15

yang

-0.15

 Preis

-0.15

yk

-0.15

yon

-0.15

yx

-0.15

ickle

-0.14

POSITIVE LOGITS

venile

0.32

arez

0.29

iced

0.24

icy

0.23

icing

0.23

illi

0.22

icer

0.22

icios

0.21

erg

0.21

ventus

0.21

Activations Density 0.004%

mentions of the word "Ju" and its variations, likely signaling references to juvenile or youth-related contexts

No Comments

No Known Activations