Pith. sign in

REVIEW 11 cited by

Active Preference Optimization for Sample Efficient RLHF

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.10500 v3 pith:KG3JFTDV submitted 2024-02-16 cs.LG cs.AIcs.CL

Active Preference Optimization for Sample Efficient RLHF

classification cs.LG cs.AIcs.CL
keywords preferencecontextsdatalowerpolicyrlhfsamplesub-optimality
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Large Language Models (LLMs) aligned using Reinforcement Learning from Human Feedback (RLHF) have shown remarkable generation abilities in numerous tasks. However, collecting high-quality human preferences creates costly bottlenecks in practical deployments, and hence, training data are often budgeted. In these scenarios, it is crucial to collect training data (e.g., contexts, a pair of generations for each context, and a preference indicating which generation is better) carefully, yet most of the existing methods sample contexts uniformly at random from a given collection. Given this, under the Bradley-Terry-Luce preference model and with a small budget of training data, we show that uniform sampling of contexts could lead to a policy (i.e., an aligned model) that suffers a constant sub-optimality gap from the optimal policy. This highlights the need for an adaptive context sampling strategy for effective alignment under a small sample budget. To address this, we reformulate RLHF within the contextual preference bandit framework, treating generations as actions, and give a nearly complete characterization of the sub-optimality gap in terms of both lower and upper bounds. First, when the action set is a $d$-dimensional hypercube and the number of samples is $T$, we show an $\Omega(d/\sqrt{T})$ lower bound. Next, we propose an algorithm, $\textit{Active Preference Optimization}$ ($\texttt{APO}$), that iteratively collects preferences for the most uncertain contexts. We show that the sub-optimality gap of the policy learned via $\texttt{APO}$ matches the lower bound up to a log factor and a non-linearity constant. Finally, we perform experiments on practical datasets to validate $\texttt{APO}$'s efficacy over existing methods, establishing it as a sample-efficient and cost-effective solution for LLM alignment.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 11 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Active Learning for Stochastic Contextual Linear Bandits

    cs.LG 2026-05 unverdicted novelty 7.0

    Active context sampling algorithm for contextual linear bandits achieves instance-dependent guarantees improving over minimax rate by up to sqrt(d) and reduces samples needed in empirical tasks.

  2. MASS-DPO: Multi-negative Active Sample Selection for Direct Policy Optimization

    cs.LG 2026-05 unverdicted novelty 7.0

    MASS-DPO derives a Plackett-Luce-specific log-determinant Fisher information objective to select non-redundant negative samples, matching or exceeding multi-negative DPO performance with substantially fewer negatives ...

  3. Beyond Variance: Prompt-Efficient RLVR via Rare-Event Amplification and Bidirectional Pairing

    cs.LG 2026-02 unverdicted novelty 7.0

    Positive-negative prompt pairing with weighted GRPO improves RLVR sample efficiency, raising AIME 2025 Pass@8 from 16.8 to 22.2 on Qwen2.5-Math-7B while matching large-scale training.

  4. Reinforcement Learning for Reasoning in Large Language Models with One Training Example

    cs.LG 2025-04 accept novelty 7.0

    One training example via RLVR boosts LLM math reasoning from 17.6% to 35.7% average across six benchmarks.

  5. Wasserstein Distributionally Robust Regret Optimization for Reinforcement Learning from Human Feedback

    cs.LG 2026-04 conditional novelty 6.5

    Wasserstein distributionally robust regret optimization for RLHF admits a promptwise water-filling solution and a GRPO-compatible sampled bonus that mitigates over-optimization better than value-robust DRO.

  6. Spectral Souping: A Unified Framework for Online Preference Alignment

    cs.LG 2026-05 unverdicted novelty 6.0

    Spectral Souping learns offline specialized policies for fine-grained preferences and merges them online using a discovered universal spectral representation for efficient LLM alignment.

  7. Wasserstein Distributionally Robust Regret Optimization for Reinforcement Learning from Human Feedback

    cs.LG 2026-04 unverdicted novelty 6.0

    DRRO for RLHF replaces worst-case value with worst-case regret in Wasserstein DRO, producing an exact water-filling solution under l1 ambiguity and a practical sampled-bonus algorithm that reduces proxy over-optimization.

  8. Wasserstein Distributionally Robust Regret Optimization for Reinforcement Learning from Human Feedback

    cs.LG 2026-04 unverdicted novelty 6.0

    DRRO for RLHF minimizes worst-case regret relative to the best policy under Wasserstein reward perturbations, yielding an exact inner solution and water-filling policy structure for the promptwise simplex model plus a...

  9. Provably avoiding over-optimization in Direct Preference Optimization without knowing the data distribution

    cs.LG 2026-02 unverdicted novelty 6.0

    PEPO uses pessimistic ensembling of DPO policies on data subsets to achieve single-policy concentrability sample bounds and avoid over-optimization in tabular settings.

  10. Provably avoiding over-optimization in Direct Preference Optimization without knowing the data distribution

    cs.LG 2026-02 unverdicted novelty 5.0

    PEPO is a single-step pessimistic ensemble algorithm for direct preference optimization that provably avoids over-optimization by depending only on single-policy concentrability without knowing the data distribution o...

  11. Reinforcement Learning from Human Feedback: A Statistical Perspective

    stat.ML 2026-04 accept novelty 2.0

    A statistical survey of RLHF for LLM alignment that connects preference learning and policy optimization to models like Bradley-Terry-Luce while reviewing methods, extensions, and open challenges.