INDEX

Explanations

quantitative data or statistical information related terms

oai_token-act-pair · gpt-3.5-turbo

phrases related to self-reported data and statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Äĩ

-0.51

 grate

-0.49

rawdownloadcloneembedreportprint

-0.49

 Mehran

-0.45

 EntityItem

-0.43

ÃĥÃĤ

-0.43

 kidding

-0.42

Wyn

-0.42

STER

-0.42

Pwr

-0.41

POSITIVE LOGITS

iod

0.52

 eternity

0.52

cycles

0.52

usterity

0.49

power

0.48

DonaldTrump

0.46

ties

0.44

arsity

0.44

wealth

0.43

 footing

0.43

Activations Density 1.928%

quantitative data or statistical information related terms

phrases related to self-reported data and statistics

No Comments

No Known Activations