INDEX

Explanations

phrases indicating moral or ethical discussions related to authority figures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

aza

-0.16

yne

-0.16

oken

-0.15

illow

-0.15

AppComponent

-0.14

ï¸ı

-0.14

ãģŁãģ¡ãģ¯

-0.14

 Pioneer

-0.14

éĹ»

-0.14

icie

-0.14

POSITIVE LOGITS

 whose

0.20

 whom

0.18

whose

0.16

who

0.16

adle

0.16

 which

0.15

 mand

0.15

Æ°á»£ng

0.14

istor

0.14

æ±

0.14

Activations Density 0.121%

phrases indicating moral or ethical discussions related to authority figures

No Comments

No Known Activations