INDEX

Explanations

references to damage and safety concerns

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 16-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L16R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.16.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

acle

-0.16

 Nack

-0.15

weg

-0.15

 Adam

-0.15

-0.14

 retr

-0.14

 Robertson

-0.14

enberg

-0.14

vek

-0.14

POSITIVE LOGITS

 damage

0.39

 Damage

0.34

damage

0.33

Damage

0.29

 DAMAGE

0.29

 damages

0.28

 damaged

0.25

_damage

0.24

amage

0.24

 damaging

0.24

Activations Density 0.119%

references to damage and safety concerns

No Known Activations