INDEX

Explanations

words related to subatomic particles called "positrons"

oai_token-act-pair · gpt-3.5-turbo

the term "rons," likely indicating a focus on particles or subatomic entities

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

aid

-0.61

 holy

-0.61

 teenager

-0.61

 Fisher

-0.60

 payroll

-0.60

ut

-0.60

 cleric

-0.58

 sting

-0.58

Hig

-0.58

 Christian

-0.58

POSITIVE LOGITS

rons

4.94

ron

1.92

RON

1.55

rals

1.47

ront

1.26

rone

1.23

ronic

1.20

rolet

1.16

rium

1.13

rogens

1.12

Activations Density 0.013%

words related to subatomic particles called "positrons"

the term "rons," likely indicating a focus on particles or subatomic entities

No Comments

No Known Activations