INDEX

Explanations

statements beginning with "Yes" and followed by affirming or agreeing words or phrases

oai_token-act-pair · gpt-3.5-turbo

affirmations or positive acknowledgments

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-+-+

-0.69

perial

-0.65

inese

-0.65

 rall

-0.62

isf

-0.62

 Gleaming

-0.61

 comprom

-0.61

kefeller

-0.60

actionDate

-0.57

tnc

-0.57

POSITIVE LOGITS

terday

1.63

hua

1.00

sir

0.94

 indeed

0.89

0.86

TER

0.83

yes

0.80

!,

0.78

ter

0.77

hur

0.76

Activations Density 0.023%

statements beginning with "Yes" and followed by affirming or agreeing words or phrases

affirmations or positive acknowledgments

No Comments

No Known Activations