INDEX

Explanations

references to the social media platform Facebook

oai_token-act-pair · gpt-3.5-turbo

references to Facebook and its features

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

theless

-0.75

 certific

-0.69

ruction

-0.68

iod

-0.68

aterasu

-0.66

Ö¼

-0.65

unta

-0.64

 negoti

-0.64

thur

-0.61

 respectively

-0.61

POSITIVE LOGITS

 Facebook

3.59

 facebook

2.78

Facebook

2.70

facebook

2.29

 Instagram

2.20

 Twitter

2.07

 LinkedIn

2.02

 Snapchat

1.99

 WhatsApp

1.98

 YouTube

1.86

Activations Density 0.028%

references to the social media platform Facebook

references to Facebook and its features

No Comments

No Known Activations