Skip to content
OpenSmartRoute

LLMs1 min read

CLLPU Benchmark Reveals Propagation Challenges in Multilingual LLM Unlearning

A new benchmark, CLLPU, assesses multilingual LLM unlearning by evaluating the extent of knowledge suppression across languages. Experiments with Llama-3.1-8B-Instruct show that universal suppression often fails to fully remove target knowledge, while language-conditioned confinement leads to uncontrolled propagation.

By OpenSmartRoute editorial · written through the router by writer-small

From arXiv cs.CL - “Beyond Cross-Lingual Transfer: Benchmarking Propagation Boundaries in Multilingual LLM Unlearning”

The research introduces CLLPU (Cross-Lingual and Language-Bound Protocol for LLM Unlearning), a benchmark designed to evaluate multilingual LLM unlearning beyond cross-lingual transfer. The benchmark utilizes two distinct settings: common-goal forgetting, aiming for suppression across all languages, and language-conditioned forgetting, restricting suppression to a specific language. CLLPU constructs 800 matched knowledge-unit pairs and 72,000 QA instances across ten languages, employing goal-guided topic pairing, schema-aware relation matching, and dual-anchor multilingual translation. Six representative methods were evaluated using Llama-3.1-8B-Instruct, revealing divergent failure modes. Specifically, universal suppression resulted in incomplete forgetting, while language-conditioned confinement caused propagation beyond the intended boundary. The research indicates that general multilingual utility can mask damage to neighboring knowledge. These findings highlight propagation control as a key challenge in multilingual LLM unlearning.

Source: https://arxiv.org/abs/2609.05976

Published Sep 9, 2026 · updated Sep 9, 2026 · 124 words

Keep reading

Related posts

More in LLMs

LLMs1 min read

AlignDiff: Filtering Preference Data with Model Signals

AlignDiff is a new framework that improves LLM alignment by prioritizing challenging preference samples based on model-intrinsic signals like negative log-likelihood gaps. Evaluations on LLaMA, Qwen, and common benchmarks demonstrate consistent performance improvements over established baselines.

LLMs1 min read

ModularPhaseNet: Discretized Phase Geometry for Transformers

ModularPhaseNet introduces a classical, integer-computable discretization of complex phase geometry, offering a new approach to Transformer architecture. It utilizes cyclic group operations for phase representation, aiming to improve semantic hierarchy and contextual consistency.

LLMs1 min read

Query Rewriting Enhances RAG Performance Through Complementarity

Research found that combining query rewriting strategies with a strong RAG baseline yields significant improvements in retrieval accuracy, primarily driven by complementarity rather than increased retrieval budget. A cost-aware router further optimizes this approach.