INDEX

Explanations

terms related to law enforcement actions and criminal activities

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

 Prosecutor

-0.16

 Victim

-0.16

victim

-0.15

 Abuse

-0.15

 Prosec

-0.15

aterno

-0.15

regor

-0.14

Ä±za

-0.14

 Discrim

-0.14

 victim

-0.14

POSITIVE LOGITS

 suspects

0.46

 accused

0.40

 suspect

0.39

cul

0.37

 suspected

0.35

 wanted

0.32

sus

0.31

 culprit

0.31

 defendants

0.30

wanted

0.28

Activations Density 0.230%

terms related to law enforcement actions and criminal activities

No Comments

No Known Activations