INDEX

Explanations

terms related to crime and deception

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 22-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L22R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.22.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

 kles

-0.14

 Carrier

-0.14

 bouquet

-0.14

 hoÃłng

-0.14

Collapsed

-0.13

Chance

-0.13

erte

-0.13

decess

-0.13

åĿ¦

-0.13

POSITIVE LOGITS

 targeting

0.28

 target

0.26

 intent

0.23

target

0.21

 targets

0.19

 attack

0.19

arget

0.18

 loose

0.18

 Target

0.18

intent

0.17

Activations Density 0.334%

terms related to crime and deception

No Comments

No Known Activations