INDEX

Explanations

references to championship titles and matches in a wrestling context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.18.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

linger

-0.15

onda

-0.14

åħ±åĴĮåĽ½

-0.14

.xyz

-0.14

opsis

-0.14

opr

-0.14

OPS

-0.14

ãĢģãĢģ

-0.13

hsi

-0.13

à¤Ĥà¤ľà¤¨

-0.13

POSITIVE LOGITS

 crown

0.40

 title

0.33

 throne

0.32

 crowned

0.30

 Crown

0.28

 coron

0.28

 crow

0.27

.title

0.26

 titles

0.25

title

0.25

Activations Density 0.144%

references to championship titles and matches in a wrestling context

No Comments

No Known Activations