INDEX

Explanations

phrases related to humor, particularly jokes

oai_token-act-pair · gpt-3.5-turbo

instances of the word "joke" and related terms or phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Releases

-0.74

obil

-0.68

antage

-0.65

Catal

-0.64

olar

-0.63

licts

-0.63

Dri

-0.60

 Seeking

-0.60

sections

-0.60

 Identified

-0.60

POSITIVE LOGITS

 joke

3.97

 jokes

2.67

 joking

2.06

 laugh

1.90

 prank

1.75

 chuckle

1.63

 humor

1.59

gag

1.54

 joked

1.52

 laughing

1.50

Activations Density 0.015%

phrases related to humor, particularly jokes

instances of the word "joke" and related terms or phrases

No Comments

No Known Activations