INDEX

Explanations

dates represented by month and day

oai_token-act-pair · gpt-3.5-turbo

the abbreviation for "number" or "numbered" as in dates, typically written with a dot (period) following "No." or "Nr."

oai_token-act-pair · gpt-4-turbo

punctuation marks, specifically periods

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

 iceberg

-0.81

 retrospective

-0.74

 transact

-0.73

 metic

-0.72

 snapshot

-0.71

osaurs

-0.71

uner

-0.68

chwitz

-0.68

¥ŀ

-0.68

omn

-0.67

POSITIVE LOGITS

 Afterwards

0.79

ItemTracker

0.72

 Akron

0.71

assetsadobe

0.69

 Eventually

0.69

 Luckily

0.67

 Speedway

0.67

Lv

0.67

Lt

0.66

Cu

0.66

Activations Density 0.287%

dates represented by month and day

the abbreviation for "number" or "numbered" as in dates, typically written with a dot (period) following "No." or "Nr."

punctuation marks, specifically periods

No Comments

No Known Activations