INDEX

Explanations

phrases indicating prompting or suggesting actions, particularly involving looking or checking

oai_token-act-pair · gpt-3.5-turbo

phrases that encourage taking a closer look or examination

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Develop

-0.67

 occupying

-0.64

ted

-0.62

eki

-0.61

 disputes

-0.60

dk

-0.60

 agreements

-0.60

omething

-0.59

ambo

-0.58

 defects

-0.58

POSITIVE LOGITS

 look

1.29

 peek

1.20

 glance

1.10

 stroll

1.10

 glimpse

1.09

 listen

1.03

 moment

1.01

 screenshot

0.93

 minute

0.92

 trip

0.91

Activations Density 0.089%

phrases indicating prompting or suggesting actions, particularly involving looking or checking

phrases that encourage taking a closer look or examination

No Comments

No Known Activations