INDEX

Explanations

mentions of terrorist organizations and related names

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

.sponge

-0.18

ç¿

-0.15

artz

-0.15

HAL

-0.15

assi

-0.15

ifndef

-0.14

ORB

-0.14

arts

-0.14

 autos

-0.14

 Äĳá»©ng

-0.13

POSITIVE LOGITS

-Qaeda

0.26

 Qaeda

0.25

aeda

0.21

queda

0.18

 Gore

0.17

hur

0.17

leg

0.16

azeera

0.16

-Q

0.16

 queda

0.16

Activations Density 0.008%

mentions of terrorist organizations and related names

No Comments

No Known Activations