Acquisition route affects forgetting rates in multimodal models, with text-pathway knowledge forgetting faster than audio-pathway knowledge in music understanding tasks.
Title resolution pending
15 Pith papers cite this work. Polarity classification is still indexing.
15
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
PROXYMIX learns a dynamic replay controller on a small proxy model and transfers it to a large target model, improving accuracy by 3.4 points and reducing forgetting by 3.5 points on LLaMA-3-8B continual tuning sequences.
LM agents' changeable modules prevent persistent identity and sanction sensitivity, making reputation mechanisms structurally inapplicable and requiring protocol-based behavioral harnesses instead.
SynLearner lets LLMs improve synthetic data generation on later tasks in a stream by learning reusable patterns and balancing quality with diversity from feedback on earlier tasks.
OP-Mix is an on-policy data mixing method that uses low-rank adapter interpolation to find near-optimal data mixtures throughout language model training with reduced compute.
SLICE applies gradient surgery via projection and truncated SVD to initialize LoRA adapters, yielding better stability-plasticity trade-offs on continual learning benchmarks including adversarial task sequences.
HypEHR is a hyperbolic embedding model for EHR data that uses Lorentzian geometry and hierarchy-aware pretraining to answer clinical questions nearly as well as large language models but with much smaller size.
Fine-tuned LLM judges struggle with future-proofing to newer generators but maintain backward-compatibility more easily; DPO training and continual learning improve adaptation while all models degrade on unseen questions.
CroCo applies English-reward-ranked self-generations for contrastive preference tuning that improves two LLMs on structured and open-ended tasks across 14 languages without language-specific annotations.
RocketSmith is an LLM-based agentic system that designs four high-powered rockets via additive manufacturing, with two achieving stable launches and recovery after reaching 80% of simulated apogee.
Multi-stage LLM training plus compiler-guided error repair boosts functional equivalence in Java-to-Cangjie translation by 6.06% over prior methods despite scarce parallel data.
Memini is introduced as a graph-based external memory using multi-timescale edge dynamics to enable emergent episodic sensitivity, consolidation, and selective forgetting in LLM systems.
This survey defines the Federated Continual Learning problem, proposes a taxonomy for approaches, reviews applications and metrics, and identifies open challenges in lifelong privacy-preserving learning on non-stationary distributed data.
Position paper calling for stronger evidentiary standards and a diagnostic checklist in anthropomorphic misalignment research.
citing papers explorer
-
When the Same Musical Knowledge Forgets Differently: A Clean Probe of Pathway-Dependent Forgetting
Acquisition route affects forgetting rates in multimodal models, with text-pathway knowledge forgetting faster than audio-pathway knowledge in music understanding tasks.
-
Dynamic Proxy-Mixing: Transferring Replay Controllers from Small to Large Models for Continual Instruction Tuning
PROXYMIX learns a dynamic replay controller on a small proxy model and transfers it to a large target model, improving accuracy by 3.4 points and reducing forgetting by 3.5 points on LLaMA-3-8B continual tuning sequences.
-
Dissociative Identity: Language Model Agents Lack Grounding for Reputation Mechanisms
LM agents' changeable modules prevent persistent identity and sanction sensitivity, making reputation mechanisms structurally inapplicable and requiring protocol-based behavioral harnesses instead.
-
Make LLM Learn to Synthesize from Streaming Experiences through Feedback
SynLearner lets LLMs improve synthetic data generation on later tasks in a stream by learning reusable patterns and balancing quality with diversity from feedback on earlier tasks.
-
Always Learning, Always Mixing: Efficient and Simple Data Mixing All The Time
OP-Mix is an on-policy data mixing method that uses low-rank adapter interpolation to find near-optimal data mixtures throughout language model training with reduced compute.
-
Low-Rank Adapters Initialization via Gradient Surgery for Continual Learning
SLICE applies gradient surgery via projection and truncated SVD to initialize LoRA adapters, yielding better stability-plasticity trade-offs on continual learning benchmarks including adversarial task sequences.
-
HypEHR: Hyperbolic Modeling of Electronic Health Records for Efficient Question Answering
HypEHR is a hyperbolic embedding model for EHR data that uses Lorentzian geometry and hierarchy-aware pretraining to answer clinical questions nearly as well as large language models but with much smaller size.
-
On the Shelf Life of Fine-Tuned LLM-Judges: Future-Proofing, Backward-Compatibility, and Question Generalization
Fine-tuned LLM judges struggle with future-proofing to newer generators but maintain backward-compatibility more easily; DPO training and continual learning improve adaptation while all models degrade on unseen questions.
-
CroCo: Cross-Lingual Contrastive Preference Tuning on Self-Generations
CroCo applies English-reward-ranked self-generations for contrastive preference tuning that improves two LLMs on structured and open-ended tasks across 14 languages without language-specific annotations.
-
RocketSmith: Agentic Additive Manufacturing of High-Powered Rockets
RocketSmith is an LLM-based agentic system that designs four high-powered rockets via additive manufacturing, with two achieving stable launches and recovery after reaching 80% of simulated apogee.
-
Boosting Automatic Java-to-Cangjie Translation with Multi-Stage LLM Training and Error Repair
Multi-stage LLM training plus compiler-guided error repair boosts functional equivalence in Java-to-Cangjie translation by 6.06% over prior methods despite scarce parallel data.
-
Continual Knowledge Updating in LLM Systems: Learning Through Multi-Timescale Memory Dynamics
Memini is introduced as a graph-based external memory using multi-timescale edge dynamics to enable emergent episodic sensitivity, consolidation, and selective forgetting in LLM systems.
-
Federated continual learning: A comprehensive survey on lifelong and privacy-preserving learning over distributed and non-stationary data
This survey defines the Federated Continual Learning problem, proposes a taxonomy for approaches, reviews applications and metrics, and identifies open challenges in lifelong privacy-preserving learning on non-stationary distributed data.
-
Position: Anthropomorphic Misalignment Research Needs Stronger Evidence
Position paper calling for stronger evidentiary standards and a diagnostic checklist in anthropomorphic misalignment research.
- Improve Large Language Model Systems with User Logs