INDEX

Explanations

phrases or words related to a specific style or design

oai_token-act-pair · gpt-3.5-turbo

references to a specific format or method of presentation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

umi

-0.78

teenth

-0.71

 Stars

-0.69

amen

-0.65

usions

-0.64

 Abraham

-0.63

mus

-0.62

ora

-0.61

azar

-0.61

å¥³

-0.61

POSITIVE LOGITS

lihood

0.91

style

0.87

heet

0.86

folios

0.82

styles

0.81

folio

0.76

 Canaver

0.68

 Dragonbound

0.67

soDeliveryDate

0.66

 prevailed

0.66

Activations Density 0.008%

phrases or words related to a specific style or design

references to a specific format or method of presentation

No Comments

No Known Activations