INDEX

Explanations

shortened words or abbreviations preceded by a specific symbol

oai_token-act-pair · gpt-3.5-turbo

the letter 'b' at various frequencies throughout the text

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

 BART

-0.70

 proceedings

-0.66

 Ãľ

-0.63

 Carney

-0.61

 territ

-0.61

 unaccount

-0.61

 essence

-0.60

 entrenched

-0.59

 subsistence

-0.58

 Stall

-0.58

POSITIVE LOGITS

odies

1.44

rief

1.35

rought

1.32

amboo

1.26

rows

1.24

inders

1.22

usting

1.21

asket

1.21

ilingual

1.20

acteria

1.20

Activations Density 0.059%

shortened words or abbreviations preceded by a specific symbol

the letter 'b' at various frequencies throughout the text

No Comments

No Known Activations