INDEX

Explanations

elements related to humor and jokes

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

indow

-0.17

 strtol

-0.15

 NavParams

-0.14

INFRINGEMENT

-0.14

shed

-0.13

.Validation

-0.13

Ð²Ð°Ð»Ð¸

-0.13

ollywood

-0.13

 Matters

-0.13

 Pron

-0.13

POSITIVE LOGITS

 joke

0.35

 jokes

0.32

 jest

0.31

 cracks

0.31

rib

0.30

-j

0.30

 practical

0.30

 cracked

0.29

 cracking

0.28

JO

0.28

Activations Density 0.145%

elements related to humor and jokes

No Comments

No Known Activations

elements related to humor and jokes

No Comments

No Known Activations