INDEX

Explanations

words related to product management or marketing

oai_token-act-pair · gpt-3.5-turbo

references to 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 dodge

-0.88

 ward

-0.78

 fighters

-0.72

 dwar

-0.72

 dodging

-0.70

dod

-0.70

 flashbacks

-0.69

 diplom

-0.68

 relay

-0.68

vet

-0.68

POSITIVE LOGITS

 Product

3.54

Product

2.36

 Products

2.18

 PRODUCT

1.72

 Customer

1.68

product

1.65

Products

1.64

 Retail

1.52

Customer

1.51

 Sales

1.40

Activations Density 0.023%

words related to product management or marketing

references to products

No Comments

No Known Activations