INDEX

Explanations

references to bombs and explosive devices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 20-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L20R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.20.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

ennen

-0.15

 knife

-0.15

ensis

-0.15

 Knife

-0.15

 blades

-0.15

licate

-0.15

-0.14

eni

-0.14

Knife

-0.14

idot

-0.14

POSITIVE LOGITS

 deton

0.32

çĤ¸

0.31

 explosive

0.28

 explodes

0.28

 exploded

0.28

çĪĨ

0.28

 explode

0.27

 Explos

0.27

 explosion

0.27

 explosives

0.27

Activations Density 0.063%

references to bombs and explosive devices

No Comments

No Known Activations