INDEX

Explanations

references to government surveillance and watchlists

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

ypi

-0.14

ckill

-0.14

sitemap

-0.14

ê²

-0.14

rir

-0.13

atta

-0.13

ettel

-0.13

richt

-0.13

ibu

-0.13

_drag

-0.13

POSITIVE LOGITS

 card

0.28

 cards

0.27

 Cards

0.23

 traveler

0.23

bi

0.23

-card

0.22

 Trusted

0.22

 Secure

0.21

 Card

0.20

.card

0.20

Activations Density 0.008%

references to government surveillance and watchlists

No Comments

No Known Activations