INDEX

Explanations

phrases related to danger or potential harm

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

iao

-0.16

agues

-0.16

ares

-0.15

arez

-0.15

.sharedInstance

-0.15

ero

-0.15

ignon

-0.14

amin

-0.14

-called

-0.14

ÐŁÐļ

-0.14

POSITIVE LOGITS

 danger

0.17

 dangers

0.17

lessly

0.17

çĬ¶

0.17

baar

0.16

elm

0.16

weigh

0.16

jsp

0.15

ĺ

0.14

ources

0.14

Activations Density 0.034%

phrases related to danger or potential harm

No Comments

No Known Activations