Formalizes continual model routing (CMR), releases CMRBench with over 2000 models, and presents CARvE which outperforms retrieval, fine-tuning and adapter-merging baselines on model/family/domain accuracy.
hub
Continual lifelong learning with neural networks: A review
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
hub tools
representative citing papers
Formalizes Reasoning Portability (RP) and proposes RDB-CL to modulate per-sample KL regularization in RLVR for MLLM continual learning, achieving +12.0% Last accuracy over vanilla RLVR baseline by preserving reusable reasoning on high-RP samples.
LoRA adapters should be scaled by 1/sqrt(rank) rather than 1/rank to stabilize learning and enable effective use of higher ranks during fine-tuning of large language models.
Heuresis evaluates six search strategies for LLM research agents and shows they steer ideas along quality-diversity-novelty axes but fail to generate novel ideas that match or exceed known high-performing recipes.
Silent collapse in recursive learning contracts internal distributions like entropy and diversity despite stable metrics, preceded by three precursors that enable the MTR monitoring framework to intervene early.
Task-aligned supervised geometric stability predicts linear steerability with high accuracy while unsupervised stability detects representational drift earlier and with lower false alarms than CKA or Procrustes.
Proposes a self-evolving cognitive framework integrating causal world modeling, intervention-driven reasoning, and continual refinement for embodied scientific intelligence.
C2FL proposes spatial clustering plus continual learning techniques inside federated learning to maintain performance under combined spatial heterogeneity and temporal drift.
SF-NorMuon is a new schedule-free spectral optimizer that closes the gap with tuned AdamW on 125M-772M parameter models across 1-8x Chinchilla horizons while providing stationarity guarantees.
IEFF enables retrain-free feature efficiency rollouts in ranking systems by elastically controlling feature coverage at serving time, achieving 5x faster rollouts, zero retraining GPU cost, and 50-55% less performance degradation than abrupt feature removal.
Proposes LoRA-based mixture-of-experts with autoencoder routing for continual bidirectional motion-language learning, reporting near-zero forgetting on a 5-task HumanML3D benchmark derived via semantic clustering.
RIZZ is a continual adaptation framework for black-box LLM agents that uses dynamically spawned memory branches, context-aware routing, verifier-gated updates, and prompt compilation to control interference across nonstationary inputs.
citing papers explorer
-
A Rank Stabilization Scaling Factor for Fine-Tuning with LoRA
LoRA adapters should be scaled by 1/sqrt(rank) rather than 1/rank to stabilize learning and enable effective use of higher ranks during fine-tuning of large language models.