INDEX

Explanations

references to the letter "N" followed by numbers

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

èıĮ

-0.15

 disturbed

-0.15

_mk

-0.14

imps

-0.14

 Consulting

-0.14

dyn

-0.13

SG

-0.13

plets

-0.13

Ä±sÄ±ndan

-0.13

POSITIVE LOGITS

iger

0.26

ollywood

0.25

igeria

0.24

aira

0.23

nam

0.20

dig

0.20

zer

0.20

ai

0.19

ger

0.19

ige

0.18

Activations Density 0.006%

references to the letter "N" followed by numbers

No Comments

No Known Activations