INDEX

Explanations

product or service names mentioned in a list of features or ingredients

oai_token-act-pair · gpt-3.5-turbo

mention of pros, ingredients, and features related to products or recipes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 externalToEVAOnly

-0.81

lit

-0.76

 visitation

-0.74

iably

-0.69

 registry

-0.68

aneously

-0.67

 inciner

-0.66

Ð¾Ð

-0.63

 payday

-0.63

 auction

-0.63

POSITIVE LOGITS

Pros

1.10

 Pros

0.98

ecause

0.86

 Features

0.84

Features

0.84

ï¸

0.76

 Finish

0.76

reads

0.76

_-

0.75

 Includes

0.74

Activations Density 0.045%

product or service names mentioned in a list of features or ingredients

mention of pros, ingredients, and features related to products or recipes

No Comments

No Known Activations