INDEX

Explanations

economic and numerical information, such as statistics, figures, and levels

oai_token-act-pair · gpt-3.5-turbo

numerical data and levels indicative of trends or thresholds

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

pora

-0.70

andals

-0.65

artifacts

-0.64

thia

-0.64

among

-0.63

behind

-0.62

 Miscellaneous

-0.61

VPN

-0.60

 Sons

-0.60

 Rings

-0.60

POSITIVE LOGITS

 threshold

1.58

 limit

1.53

 thresholds

1.41

 cutoff

1.33

 standards

1.29

 limits

1.26

 norms

1.24

 estimate

1.23

 guideline

1.22

 maximum

1.22

Activations Density 0.404%

economic and numerical information, such as statistics, figures, and levels

numerical data and levels indicative of trends or thresholds

No Comments

No Known Activations