INDEX

Explanations

proper nouns, potentially names of people or companies

oai_token-act-pair · gpt-3.5-turbo

instances of the term "Ad" that likely refer to advertisements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

-0.83

ANGE

-0.76

externalActionCode

-0.75

ãĥīãĥ©ãĤ´ãĥ³

-0.73

-0.67

OUP

-0.67

 IPCC

-0.65

 biomass

-0.64

OURCE

-0.64

 Weir

-0.63

POSITIVE LOGITS

vertising

1.45

elaide

1.29

vertis

1.29

vance

1.24

vantage

1.18

vertisements

1.14

mittedly

1.06

mir

1.02

ventures

1.01

ept

1.00

Activations Density 0.008%

proper nouns, potentially names of people or companies

instances of the term "Ad" that likely refer to advertisements

No Comments

No Known Activations