INDEX

Explanations

references to legal authorities and officials, particularly pertaining to government investigations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

ssize

-0.15

 Angiosper

-0.15

 burg

-0.14

lopedia

-0.14

geb

-0.14

NES

-0.14

ANNEL

-0.14

abella

-0.14

 Tribunal

-0.14

ghi

-0.14

POSITIVE LOGITS

AG

0.31

DOJ

0.30

 Justice

0.29

 Attorney

0.28

AG

0.26

Att

0.25

 Department

0.24

Justice

0.24

Att

0.23

att

0.23

Activations Density 0.072%

references to legal authorities and officials, particularly pertaining to government investigations

No Comments

No Known Activations