INDEX

Explanations

phrases that indicate possession or ownership

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

imson

-0.18

won

-0.15

ateful

-0.15

 eldre

-0.14

elin

-0.14

 ktÃ³

-0.14

ocrates

-0.14

 Ð¿Ð¾ÑıÐ²Ð¸

-0.14

fil

-0.14

ivre

-0.13

POSITIVE LOGITS

 always

0.35

always

0.31

 Always

0.29

Always

0.29

 ALWAYS

0.25

 long

0.24

 historically

0.24

 traditionally

0.23

 tended

0.22

 sempre

0.20

Activations Density 0.178%

phrases that indicate possession or ownership

No Comments

No Known Activations