INDEX

Explanations

mentions of locations or spaces that are unoccupied or available

oai_token-act-pair · gpt-3.5-turbo

terms related to emptiness or unoccupied spaces

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

ads

-0.77

bringer

-0.73

 Protocol

-0.68

 Mother

-0.63

anners

-0.63

jet

-0.62

ather

-0.61

arta

-0.60

DK

-0.60

bit

-0.60

POSITIVE LOGITS

 vacant

3.87

 vacancy

2.59

 vacancies

2.19

 vacated

2.11

 dormant

1.25

Vac

1.22

 occupied

1.22

vac

1.22

 unused

1.19

 renovated

1.19

Activations Density 0.017%

mentions of locations or spaces that are unoccupied or available

terms related to emptiness or unoccupied spaces

No Comments

No Known Activations