INDEX

Explanations

names of individuals and characters in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

accent

-0.15

avanaugh

-0.13

inese

-0.13

burger

-0.13

 oprav

-0.13

wort

-0.13

oze

-0.13

 Arts

-0.13

ioneer

-0.13

.FindAsync

-0.13

POSITIVE LOGITS

/Dk

0.17

 Lara

0.17

THR

0.15

ãģ£ãģ

0.15

åľº

0.15

áº¡nh

0.14

/DTD

0.14

ìĿ

0.14

 sÄ©

0.14

ple

0.14

Activations Density 0.114%

names of individuals and characters in various contexts

No Comments

No Known Activations