INDEX

Explanations

dates or timestamps from online posts or publications

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Posted" followed by numbers, indicating dates and times

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

atern

-0.81

fell

-0.81

isky

-0.80

ugal

-0.75

ishi

-0.73

oshenko

-0.73

aternity

-0.72

ppard

-0.72

ivo

-0.71

$$$$

-0.71

POSITIVE LOGITS

Posted

0.83

erick

0.75

 Thumbnails

0.73

monton

0.68

 Comments

0.68

itors

0.66

Kiw

0.64

 Parade

0.64

vertis

0.64

Sun

0.63

Activations Density 0.010%

dates or timestamps from online posts or publications

instances of the word "Posted" followed by numbers, indicating dates and times

No Comments

No Known Activations