INDEX

Explanations

phrases where the word "ORE" is mentioned along with a high activation value of 9 or 10

oai_token-act-pair · gpt-3.5-turbo

the brand "ORE."

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

sid

-0.72

 angels

-0.66

st

-0.64

 Bast

-0.63

 Avalanche

-0.62

 speeding

-0.59

 Idlib

-0.59

Bec

-0.58

qu

-0.58

 retrospect

-0.58

POSITIVE LOGITS

ORE

4.42

ores

2.18

ore

2.11

ORED

1.96

ored

1.49

oring

1.46

OR

1.39

ORY

1.37

ALE

1.35

orer

1.23

Activations Density 0.008%

phrases where the word "ORE" is mentioned along with a high activation value of 9 or 10

the brand "ORE."

No Comments

No Known Activations