INDEX

Explanations

time-related durations expressed in weeks

oai_token-act-pair · gpt-3.5-turbo

references to durations of time, specifically in weeks

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

jri

-0.72

 constitu

-0.70

 culp

-0.64

ifice

-0.62

 superflu

-0.61

Cir

-0.60

 racially

-0.60

 misunderstanding

-0.60

EStream

-0.60

 cogn

-0.59

POSITIVE LOGITS

days

1.19

iversary

1.07

DAY

1.02

day

0.98

iday

0.96

week

0.96

andowski

0.95

night

0.95

orrow

0.93

uty

0.91

Activations Density 0.010%

time-related durations expressed in weeks

references to durations of time, specifically in weeks

No Comments

No Known Activations