arXiv:2607. 20438v1 Announce Type: cross Abstract: Preference-based post-training is usually understood through endpoint behavior, yet the learned update that produces this behavior remains largely opaque.
arXiv:2607. 19331v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) is rapidly advancing the reasoning capabilities of language models, yet the optimization layer that converts reward feedback into weight-space updates remains poorly understood.
arXiv:2607. 07719v1 Announce Type: cross Abstract: Parameter-efficient fine-tuning adapts a large language model to one task cheaply, but across a task sequence LoRA-style methods keep stacking low-rank updates on the same frozen weight, so each new task tends to overwrite the previous ones.
arXiv:2602. 03846v2 Announce Type: replace-cross Abstract: We develop a continual learning method for pretrained models that \emph{requires no access to old-task data}, addressing a practical barrier in foundation model adaptation where pretraining distributions are often unavailable.
arXiv:2606. 19744v1 Announce Type: cross Abstract: Aligning language models with human preferences often requires optimising multiple behavioural objectives.
arXiv:2606. 12883v1 Announce Type: new Abstract: In Low-Rank Adaptation (LoRA), the scaling factor $\alpha$ is often treated as a mere complement to the learning rate, yet its role in optimization remains poorly understood.
arXiv:2607. 28669v1 Announce Type: new Abstract: We present LARA (Lightweight Additive Residual Adaptation), a method for efficient adaptation that operates in the residual stream of a frozen model rather than in its weights.
arXiv:2605. 07284v2 Announce Type: replace Abstract: A late-layer change learned during post-training may work on the base model's earlier state, or it may depend on earlier computation learned with it.
In Low-Rank Adaptation (LoRA), the scaling factor $α$ is often treated as a mere complement to the learning rate, yet its role in optimization remains poorly understood. In this paper, we reveal that the scaling factor $α$ and the learning rate function differently, with $α$ emerging as the dominant driver of effective optimization, delivering gains that cannot be replicated by learning rate scaling alone.
arXiv:2601. 13020v2 Announce Type: replace-cross Abstract: Continual instruction tuning (CIT) requires multimodal large language models (MLLMs) to adapt to a stream of tasks without forgetting prior capabilities.
arXiv:2608. 12332v1 Announce Type: cross Abstract: In recent years, low-rank adaptation (LoRA) has emerged as a significant paradigm that freezes pre-trained weights and introduces small, learnable adapters instead of fine-tuning the full set of parameters.
arXiv:2608. 06216v1 Announce Type: cross Abstract: Classical continual learning (CL) has primarily focused on enabling models to update and retain knowledge through parameter-centric mechanisms, e.
arXiv:2607. 05300v1 Announce Type: new Abstract: Parameter-efficient fine-tuning still leaves a broad space of behavior-changing updates reachable, so a poisoned objective can be represented and optimized.