cs.AIAug 20, 2026

Catching the Rug: Early Prediction of Fraudulent Memecoins on Solana via Machine Learning

Authors: Jianghai Li, Pavel Kuznetsov, Yury Yanovich, Konstantin Nott-Whaley, Igor Vodolazov

Organizations: Higher School of Economics Moscow, Russia · Moscow State University Moscow, Russia · Skolkovo Institute of Science and Technology Moscow, Russia · Independent Researcher London, UK · 

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

The rapid proliferation of memecoins on blockchain platforms has increased the risk of fraudulent activities, particularly rug pulls. While previous studies have focused on Ethereum-based tokens, this paper shifts the spotlight to Solana, the leading blockchain for memecoins by trading volume and token count. Unlike Ethereum, where rug pulls often exploit smart contract backdoors, Solana memecoin rug pulls are predominantly driven by liquidity manipulation and social dynamics. This research pioneers large-scale rug pull early detection in the Solana ecosystem by assembling a dataset of 6.4 million tokens over 7 months. Market analysis reveals that a vast majority of these memecoins exhibit rug pull characteristics within one hour of launch, highlighting the urgency of short-horizon prediction. Despite the absence of code-level features, we demonstrate that classic machine learning models, particularly Gradient Boosting (XGBoost), achieve robust performance in detecting potential rug pulls using only the first 5 minutes of trading data. Furthermore, we evaluate cross-platform generalization between PumpFun and Raydium, revealing that multi-source data fusion significantly mitigates domain shift and improves detection reliability. This study advances the understanding of DeFi fraud on high-throughput chains and provides a practical framework for protecting investors.

Figures & tables

PumpFunRaydium
block_timeslotslot
tx_idxcreatortx_idx
namesymboladdress
urlmintsignature
bundle_sizegas_usedblock_time
amount_of_instructionsamount_of_lookup_readssigning_wallet
TABLE I : Data columns extracted from Raydium and PumpFun.
DEXNumber of tokensTotal Swap Transactions
Raydium979655.923951×1085.923951\times 10^{8}
PumpFun63042354.226951×1084.226951\times 10^{8}
TABLE II : Dataset statistics for Raydium and PumpFun.
FeatureDescription
count_txTotal number of transactions associated with the token.
purchase_percentagePercentage of transactions classified as token purchases.
sale_percentagePercentage of transactions classified as token sales.
unique_buyersNumber of unique wallets that purchased the token.
unique_sellersNumber of unique wallets that sold the token.
total_sol_valueTotal value of transactions in SOL associated with the token.
TABLE III : Calculated features to observe token activity.
Fig. 3 : The Memecoin Future Prediction Framework uses Time tt to predict t+t0t+t_{0} .
MetricPumpfun (5mins)Raydium (5mins)
Total Transactions2.47×1082.47\times 10^{8}3.42×1073.42\times 10^{7}
Avg Tx3.93×1013.93\times 10^{1}5.88×1025.88\times 10^{2}
Median Tx1.20×1011.20\times 10^{1}3.85×1023.85\times 10^{2}
Total Volume (SOL)1.28×10171.28\times 10^{17}3.08×10163.08\times 10^{16}
Avg Volume2.04×10102.04\times 10^{10}5.29×10115.29\times 10^{11}
Avg Buyers1.83×1011.83\times 10^{1}1.54×1021.54\times 10^{2}
TABLE IV : Descriptive statistics comparison between Pumpfun and Raydium.
MetricRug PullNon Rug Pull
Pumpfun438354383597119711
Raydium2931946293194618935391893539
TABLE V : Label distribution of the testset between Pumpfun and Raydium.
Fig. 4 : Distribution of rug pull labels on Pumpfun and Raydium.
MethodTVL OR Idle
RR →\rightarrow PPP →\rightarrow RR ∪\cup PF →\rightarrow PF →\rightarrow R
Random Forest
F1(1)0.75240.72070.77850.61060.78380.78770.7408
MCC0.2874-0.04340.3558-0.24330.39380.39420.2246
AUCPRC0.75780.53030.76340.54350.78970.79780.6040
XGBoost
TABLE VI : Rug Pull Prediction Results (TVL OR Idle). TVL OR Idle denotes rug pulls identified by either 99% TVL drop OR 80% idle time. Best results per column are bolded .
More in this topic

Financial Fraud Detection

Financial fraud detection identifies suspicious transactions, accounts, claims, and behavior to prevent losses and guide investigation.

See all 137 papers

Explore similar work

CardsList

  1. Early Rug Pull Warning for BSC Meme Tokens via Multi-Granularity Wash-Trading Pattern Profiling

    Mar 14, 2026Dingding Cao, Bianbian Jiao, Jingzong Yang +2Financial Fraud DetectionDecentralized Finance

    School of Big Data, Baoshan University, Baoshan 678000, Yunnan, China · School of Information Science and Electrical Engineering, Shandong Jiaotong University, Jinan, Shandong, China

  2. Fraud Detection in Cryptocurrency Markets with Spatio-Temporal Graph Neural Networks

    Apr 27, 2026Lidia Losavio, Luca Persia, Madan Sathe +1Financial Fraud DetectionSpatiotemporal GNNs

    Faculty of Informatics, Universit`a della Svizzera italiana, Lugano, Switzerland · School of Management and Law, Zurich University of Applied Sciences, Zurich, Switzerland · Forensics and Financial Crime, Deloitte AG, Switzerland

  3. DeFiFusion: Combining Transaction Events with Smart Contracts to Detect Price Manipulation Attacks

    Sep 10, 2026Rui Cao, Shaojing Fan, Liming Fang +3Financial Fraud DetectionDecentralized Finance

    School of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, China · Department of Electrical and Computer Engineering (ECE), National University of Singapore, Singapore · School of Aerospace Engineering, Nanjing University of Aeronautics and Astronautics, China +2