INDEX

Explanations

specific items such as food, drinks, and clothing

oai_token-act-pair · gpt-3.5-turbo

references to specific food and beverage items or products

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

 analysts

-0.89

 administration

-0.83

 policymakers

-0.80

 administrators

-0.79

 counterterrorism

-0.79

 discredited

-0.79

 jurisdictions

-0.79

 administrations

-0.78

 Administration

-0.78

 analyst

-0.77

POSITIVE LOGITS

 adorable

1.16

 cute

1.15

 gorgeous

1.10

Awesome

1.07

 lovely

1.03

etsy

1.00

Mini

0.96

 Lovely

0.95

;)

0.94

 haha

0.94

Activations Density 0.520%

specific items such as food, drinks, and clothing

references to specific food and beverage items or products

No Comments

No Known Activations