INDEX

Explanations

terms related to specific industries or hobbies, such as baseball, music, hockey, cinema, soccer, craft beer, comics, rugby, journalism, animation, and gaming

oai_token-act-pair · gpt-3.5-turbo

references to various sports, forms of entertainment, and artistic mediums

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 clot

-0.69

ortium

-0.68

 Stab

-0.67

xit

-0.67

thens

-0.66

 Chow

-0.64

ixture

-0.62

ttp

-0.60

raints

-0.60

imura

-0.59

POSITIVE LOGITS

 journalism

1.01

 royalty

0.94

 storytelling

0.87

 fandom

0.87

 nerds

0.86

 careers

0.83

 circles

0.83

 enthusi

0.82

 academia

0.81

 nowadays

0.81

Activations Density 0.273%

terms related to specific industries or hobbies, such as baseball, music, hockey, cinema, soccer, craft beer, comics, rugby, journalism, animation, and gaming

references to various sports, forms of entertainment, and artistic mediums

No Known Activations