INDEX

Explanations

references to time measurements in terms of hours

oai_token-act-pair · gpt-3.5-turbo

references to hours worked or time spent

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

BuyableInstoreAndOnline

-0.63

dinand

-0.62

estine

-0.62

eers

-0.60

imaru

-0.59

untled

-0.56

 innocence

-0.56

 Instit

-0.56

advertising

-0.56

LECT

-0.54

POSITIVE LOGITS

glass

1.30

long

0.83

Ago

0.77

shift

0.74

 commute

0.73

 logged

0.73

shire

0.71

OND

0.71

work

0.71

leneck

0.70

Activations Density 0.056%

references to time measurements in terms of hours

references to hours worked or time spent

No Comments

No Known Activations