INDEX

Explanations

instances of the word "think" in various forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.18.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

ez

-0.17

fred

-0.15

å§¿

-0.15

alent

-0.15

iser

-0.15

asar

-0.14

sher

-0.14

anÃ¡

-0.14

)((((

-0.14

/by

-0.14

POSITIVE LOGITS

 about

0.24

 twice

0.23

 Twice

0.22

è±¡

0.18

tank

0.18

_about

0.18

.about

0.17

-about

0.17

 About

0.16

cape

0.16

Activations Density 0.084%

instances of the word "think" in various forms

No Comments

No Known Activations