INDEX

Explanations

phrases indicating optimization or maximizing efficiency

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

nten

-0.15

LEMENT

-0.15

-0.14

Ã©gor

-0.14

 Ð¿Ð°Ð»ÑĮ

-0.14

ilestone

-0.14

 OMAP

-0.14

 Stadt

-0.13

uali

-0.13

visor

-0.13

POSITIVE LOGITS

 maximizing

0.16

 talents

0.16

yar

0.16

 maximize

0.15

 Ð²Ð¾Ð·Ð¼Ð¾Ð¶

0.15

=max

0.15

 moÅ¾nÃ¡

0.14

 maxim

0.14

 Maxim

0.13

 Conc

0.13

Activations Density 0.066%

phrases indicating optimization or maximizing efficiency

No Comments

No Known Activations