INDEX

Explanations

information about premium services, especially in relation to insurance and subscriptions

oai_token-act-pair · gpt-3.5-turbo

references to premium services or products

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

dy

-0.81

 swat

-0.78

lyn

-0.78

 persecut

-0.74

 violent

-0.73

sy

-0.70

 violence

-0.70

 Lynch

-0.69

 mach

-0.69

Murray

-0.69

POSITIVE LOGITS

 Premium

3.83

Premium

2.34

 Platinum

1.17

 premium

1.11

 Deluxe

1.10

prem

1.03

 Insurance

0.99

 Moder

0.99

 Certification

0.98

 Flavoring

0.97

Activations Density 0.017%

information about premium services, especially in relation to insurance and subscriptions

references to premium services or products

No Comments

No Known Activations