INDEX

Explanations

references to specific terrorist organizations, particularly al-Qaeda and its affiliates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

-0.17

azo

-0.15

alborg

-0.14

assi

-0.14

.sdk

-0.14

Href

-0.13

 Homer

-0.13

 Knox

-0.13

ODY

-0.13

ARTH

-0.13

POSITIVE LOGITS

 Qaeda

0.25

-Qaeda

0.25

al

0.23

aeda

0.20

queda

0.20

 Nurs

0.19

 Fate

0.19

 ÐĹÐ°Ð²

0.19

_CI

0.16

Bay

0.16

Activations Density 0.014%

references to specific terrorist organizations, particularly al-Qaeda and its affiliates

No Comments

No Known Activations