INDEX

Explanations

spoken or written content that is physically written or displayed on something

oai_token-act-pair · gpt-3.5-turbo

phrases indicating markings or designs on objects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ealous

-0.70

quickShipAvailable

-0.69

cong

-0.68

ORTS

-0.67

CDC

-0.67

Consumer

-0.66

================================

-0.66

EH

-0.66

Independent

-0.65

conom

-0.65

POSITIVE LOGITS

 etched

0.98

 parchment

0.84

 canvas

0.84

 billboards

0.82

 toast

0.76

 posters

0.75

 inscribed

0.75

 wallpaper

0.74

 forehead

0.73

 pages

0.72

Activations Density 0.298%

spoken or written content that is physically written or displayed on something

phrases indicating markings or designs on objects

No Comments

No Known Activations