INDEX

Explanations

phrases related to negative or critical situations

oai_token-act-pair · gpt-3.5-turbo

terms related to crises, legal issues, and medical emergencies

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

Not in Any Lists

No Comments

Negative Logits

 Logo

-0.70

 logo

-0.70

 reprinted

-0.64

 Consumers

-0.62

 handwritten

-0.60

Gem

-0.59

 handmade

-0.58

 keyword

-0.58

 Vide

-0.58

 reproduced

-0.58

POSITIVE LOGITS

 mode

1.46

 limbo

1.34

livion

1.14

 obscurity

1.09

 territory

1.05

 positions

1.02

mode

1.01

 infancy

1.00

 custody

0.98

 Mode

0.96

Activations Density 0.585%

phrases related to negative or critical situations

terms related to crises, legal issues, and medical emergencies

No Comments

No Known Activations