PPL (Retrieval-augmented generation): superseded — cited as a baseline and beaten by newer methods. 0 paper(s) critique it, 2 beat it on benchmarks — #105 of 1179 most-superseded. Sub-problem: cluster led by RobustRAG. Newer alternatives in the same sub-problem include CORDON-MAS, BiRD, CleanBase, Beyond Factual Grounding, RAGShield.

Retrieval-augmented generation

PPL

Superseded baseline#105 of 1,179 most-superseded

Superseded — cited as a baseline and beaten by newer methods

0 papers critique it · 2 beat it on benchmarks

Beaten on benchmarks

Head-to-head results where a newer method reports beating PPL. Values are copied from the source paper's tables — verify against the cited paper.

alg beats PPL
0.99 vs 0.57
DACC · [NQ dataset / PRAGB attack]
Who Taught the Lie? Responsibility Attribution for Poisoned Knowledge in Retrieval-Augmented Generation
GMTP beats PPL
1.0 vs 0.996
FR (Filtering Rate) · [PoisonedRAG attack, Retrieval phase, NQ dataset]
Safeguarding RAG Pipelines with GMTP: A Gradient-based Masked Token Probability Method for Poisoned Document Detection

What to use instead

Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.

CORDON-MASCordon-MAS: Defending RAG against Knowledge Poisoning via Information-Flow Control
May 26, 2026
BiRDBiRD: A Bidirectional Ranking Defense Mechanism for Retrieval Augmented Generation
May 19, 2026
CleanBaseCleanBase: Detecting Malicious Documents in RAG Knowledge Databases
May 1, 2026
Beyond Factual GroundingBeyond Factual Grounding: The Case for Opinion-Aware Retrieval-Augmented Generation
Apr 13, 2026
RAGShieldRAGShield: Provenance-Verified Defense-in-Depth Against Knowledge Base Poisoning in Government Retrieval-Augmented Generation Systems
Apr 1, 2026
ProGRankProGRank: Probe-Gradient Reranking to Defend Dense-Retriever RAG from Corpus Poisoning
Mar 24, 2026
OpenDecoderOpenDecoder: Open Large Language Model Decoding to Incorporate Document Quality in RAG
Jan 13, 2026
SeCon-RAGSeCon-RAG: A Two-Stage Semantic Filtering and Conflict-Free Framework for Trustworthy RAG
Oct 10, 2025
RADARRADAR: A Risk-Aware Dynamic Multi-Agent Framework for LLM Safety Evaluation via Role-Specialized Collaboration
Sep 28, 2025
RAGOriginWho Taught the Lie? Responsibility Attribution for Poisoned Knowledge in Retrieval-Augmented Generation
Sep 17, 2025