INDEX

Explanations

numerical values

oai_token-act-pair · gpt-3.5-turbo

references to median values or averages related to statistics

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

witch

-0.78

Ãī

-0.71

 Explos

-0.68

 Travels

-0.65

 Operation

-0.64

Touch

-0.62

 Sirius

-0.61

FK

-0.60

Bring

-0.60

Sir

-0.60

POSITIVE LOGITS

 median

3.89

 Median

2.23

 average

1.63

 averages

1.34

avg

1.32

average

1.30

Average

1.30

 percentile

1.27

mean

1.23

 mean

1.13

Activations Density 0.013%

numerical values

references to median values or averages related to statistics

No Comments

No Known Activations