INDEX

Explanations

references to danger or hazardous situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 10-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L10R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.10.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

enance

-0.16

 noqa

-0.16

erez

-0.16

å±

-0.15

arity

-0.15

CALE

-0.15

NEY

-0.15

éĻµ

-0.14

ting

-0.14

verts

-0.14

POSITIVE LOGITS

-danger

0.20

 dangerous

0.17

 unsafe

0.16

ously

0.16

danger

0.15

vous

0.15

unsafe

0.15

 dangers

0.14

åı£

0.14

 Dangerous

0.14

Activations Density 0.037%

references to danger or hazardous situations

No Comments

No Known Activations