INDEX

Explanations

references to historical atrocities and war crimes

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

èĦ

-0.14

strike

-0.14

 ÎĳÎ³Î³

-0.14

 strike

-0.14

 Danger

-0.14

 Strike

-0.14

AGMENT

-0.14

 menace

-0.13

chin

-0.13

ÙĪØ¨ÛĮ

-0.13

POSITIVE LOGITS

gen

0.37

 genocide

0.35

 gÃ©n

0.30

/gen

0.29

 exter

0.28

 mass

0.28

gen

0.26

Gen

0.26

-gen

0.25

ocide

0.24

Activations Density 0.137%

references to historical atrocities and war crimes

No Comments

No Known Activations