INDEX

Explanations

references to legal and criminal justice terms or roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 23-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L23R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.23.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

 kÃ¥

-0.16

etin

-0.15

ë§Ĳ

-0.15

omi

-0.14

 Sabha

-0.14

chez

-0.14

/Gate

-0.14

IVA

-0.14

abyrinth

-0.14

tÄĽ

-0.14

POSITIVE LOGITS

 prosecutor

0.31

 prosecutors

0.30

 district

0.29

 District

0.27

 Assistant

0.26

 procure

0.26

 grand

0.25

 Ð¿ÑĢÐ¾Ðº

0.25

 Crown

0.25

Prosec

0.25

Activations Density 0.172%

references to legal and criminal justice terms or roles

No Comments

No Known Activations

references to legal and criminal justice terms or roles

No Comments

No Known Activations