INDEX

Explanations

elements related to the entertainment value and accessibility of films or reading materials

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

Negative Logits

YNC

-0.15

Æ°

-0.14

unction

-0.14

.uml

-0.14

ynam

-0.14

Ø¨Ø§Ø±

-0.14

edor

-0.13

á»¥

-0.13

zeichnet

-0.13

orsche

-0.13

POSITIVE LOGITS

esc

0.33

 entertainment

0.29

 diversion

0.28

 divert

0.25

 recommendation

0.25

 guilty

0.24

recommended

0.23

 escap

0.23

 recommended

0.23

 escape

0.23

Activations Density 0.157%

elements related to the entertainment value and accessibility of films or reading materials

No Known Activations