INDEX

Explanations

numerical references such as specific ages, years, and rankings

oai_token-act-pair · gpt-3.5-turbo

references to specific demographics or statistical data

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

},{"

-0.76

ugh

-0.76

nosis

-0.74

maxwell

-0.72

infeld

-0.71

 Dragonbound

-0.68

omy

-0.68

OULD

-0.66

uts

-0.65

edIn

-0.65

POSITIVE LOGITS

 being

1.34

 needing

1.33

 appearing

1.32

 becoming

1.30

 seeming

1.30

 gaining

1.28

 preferring

1.27

 having

1.26

 estimating

1.22

 concluding

1.20

Activations Density 0.297%

numerical references such as specific ages, years, and rankings

references to specific demographics or statistical data

No Comments

No Known Activations