INDEX

Explanations

time durations expressed in minutes

oai_token-act-pair · gpt-3.5-turbo

expressions of time durations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 inher

-0.68

 literature

-0.65

 trans

-0.65

lev

-0.64

 depend

-0.64

 ships

-0.62

 hiber

-0.62

ados

-0.62

 hunting

-0.61

 support

-0.60

POSITIVE LOGITS

minute

3.56

 minute

1.96

hour

1.86

second

1.83

 Minute

1.71

 minutes

1.54

mile

1.48

 Minutes

1.41

seconds

1.39

Hour

1.33

Activations Density 0.009%

time durations expressed in minutes

expressions of time durations

No Comments

No Known Activations