SABER: Switchable and Balanced Training for Efficient LLM Reasoning
Large language models (LLMs) empowered by chain-of-thought reasoning have achieved impressive accuracy on complex tasks but suffer from excessive inference costs and latency when applied uniformly to all problems. We propose SABER (Switchable and Balanced Training for Efficient LLM Reasoning), a reinforcement learning framework that endows LLMs with user-controllable, token-budgeted reasoning. SABER first profiles each training example's base-model thinking token usage and assigns it to one of the predefined budget tiers. During fine-tuning, the model is guided by system prompts and length-aware rewards to respect its assigned budget. In parallel, we incorporate no-think examples to ensure the model remains reliable even when explicit reasoning is turned off. SABER further supports four discrete inference modes - NoThink, FastThink, CoreThink, and DeepThink, enabling flexible trade-offs between latency and reasoning depth. Extensive evaluations on math reasoning (MATH, GSM8K), code generation (MBPP), and logical reasoning (LiveBench-Reasoning) demonstrate that SABER achieves high accuracy under tight budgets, graceful degradation, and effective cross-scale and cross-domain generalization. In particular, SABER-FastThink cuts reasoning length by 65.4% and yields a 3.6% accuracy gain compared with the base model on the MATH benchmark.
Code (0)
등록된 구현이 없습니다.
Tasks
Reinforcement LearningDomain GeneralizationLogical ReasoningCode GenerationSimilar Papers 제목 키워드 기반
SABER: Stability-Aware Early Exit for LLM Reasoning via Adversarial Branch Probing
Large Reasoning Models (LRMs) achieve strong reasoning capabilities, yet long-chain reasoning becomes inefficient once the intermediate answer stabilizes across reasoning steps: additional reasoning yields little margina…
Integrating Fine-Grained Audio-Visual Evidence for Robust Multimodal Emotion Reasoning
Multimodal emotion analysis is shifting from static classification to generative reasoning. Beyond simple label prediction, robust affective reasoning must synthesize fine-grained signals such as facial micro-expressions…
Multimodal ReasoningTrust or Abstain? A Self-Aware RAG Approach
Retrieval-augmented generation (RAG) improves large language models (LLMs) by incorporating external evidence, but it also introduces knowledge conflicts when retrieved contextual knowledge (CK) and parametric knowledge …
Investigating associative, switchable and negatable Winograd items on renewed French data sets
The Winograd Schema Challenge (WSC) consists of a set of anaphora resolution problems resolvable only by reasoning about world knowledge. This article describes the update of the existing French data set and the creation…
NegationWorld KnowledgeSABER: A Scalable Action-Based Embodied Dataset for Real-World VLA Adaptation
Robotic deployment in real-world environments depends on rich, domain-specific action data as much as on strong model architecture. General-purpose robot foundation models show modest performance in complex unseen tasks …