INDEX

Explanations

phrases indicating intentions or actions related to plans or decisions

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

New Auto-Interp

Configuration

fnlp/Llama3_1-8B-Base-LXM-8x/Llama3_1-8B-Base-L22M-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.22.hook_mlp_out

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

 IonicModule

-0.15

clair

-0.15

iology

-0.15

ials

-0.14

preh

-0.14

olid

-0.14

 stab

-0.14

ãģĤãģ®

-0.14

nage

-0.14

mage

-0.13

POSITIVE LOGITS

orie

0.16

uba

0.16

.beta

0.15

innacle

0.14

oba

0.14

wort

0.14

ARA

0.13

 BLOCK

0.13

Ø¯Ø§

0.13

 exit

0.13

Activations Density 0.125%

phrases indicating intentions or actions related to plans or decisions

No Comments

No Known Activations