INDEX

Explanations

references to royalty or imperial figures

oai_token-act-pair · gpt-3.5-turbo

references to emperors

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

 rental

-0.72

BIL

-0.68

LOC

-0.67

swing

-0.65

 Cheryl

-0.64

vers

-0.64

 canv

-0.63

cas

-0.63

 pickup

-0.62

 rented

-0.62

POSITIVE LOGITS

 Emperor

3.76

 emperor

2.95

 Empress

2.35

peror

2.28

perors

1.57

 Napoleon

1.41

 Buddha

1.39

 Empire

1.35

 Pope

1.33

 Vader

1.32

Activations Density 0.023%

references to royalty or imperial figures

references to emperors

No Comments

No Known Activations