INDEX

Explanations

phrases and terms that indicate widely accepted principles, methods, or guidelines

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 5-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L5R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.5.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

 Acres

-0.17

atches

-0.16

eut

-0.16

Ð»ÑıÑħ

-0.16

lis

-0.15

rÃ¡t

-0.15

åĲ¾

-0.15

fce

-0.14

 Atmos

-0.14

Ã¶t

-0.13

POSITIVE LOGITS

elden

0.15

ixe

0.14

ink

0.14

ãĥĨãĥ«

0.14

ungan

0.14

Pap

0.14

pap

0.14

Periph

0.13

imports

0.13

ily

0.13

Activations Density 0.268%

phrases and terms that indicate widely accepted principles, methods, or guidelines

No Comments

No Known Activations