INDEX

Explanations

phrases that indicate privacy protection and data handling

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

aptop

-0.16

issing

-0.16

lus

-0.16

 almost

-0.14

uzey

-0.14

osite

-0.14

 bastante

-0.14

innie

-0.14

Ð°Ð¶

-0.14

 sometimes

-0.14

POSITIVE LOGITS

nor

0.38

nor

0.33

 EVER

0.25

Nor

0.24

Nor

0.23

NOR

0.22

 ever

0.21

 knowingly

0.19

ä¹Łä¸į

0.18

-ever

0.17

Activations Density 0.251%

phrases that indicate privacy protection and data handling

No Comments

No Known Activations