Pith. sign in

REVIEW 25 cited by

Token-level Direct Preference Optimization

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 2404.11999 v5 pith:BRHDIKEQ submitted 2024-04-18 cs.CL cs.AI

Token-level Direct Preference Optimization

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

Fine-tuning pre-trained Large Language Models (LLMs) is essential to align them with human values and intentions. This process often utilizes methods like pairwise comparisons and KL divergence against a reference LLM, focusing on the evaluation of full answers generated by the models. However, the generation of these responses occurs in a token level, following a sequential, auto-regressive fashion. In this paper, we introduce Token-level Direct Preference Optimization (TDPO), a novel approach to align LLMs with human preferences by optimizing policy at the token level. Unlike previous methods, which face challenges in divergence efficiency, TDPO incorporates forward KL divergence constraints for each token, improving alignment and diversity. Utilizing the Bradley-Terry model for a token-based reward system, TDPO enhances the regulation of KL divergence, while preserving simplicity without the need for explicit reward modeling. Experimental results across various text tasks demonstrate TDPO's superior performance in balancing alignment with generation diversity. Notably, fine-tuning with TDPO strikes a better balance than DPO in the controlled sentiment generation and single-turn dialogue datasets, and significantly improves the quality of generated responses compared to both DPO and PPO-based RLHF methods. Our code is open-sourced at https://github.com/Vance0124/Token-level-Direct-Preference-Optimization.

discussion (0)

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

Forward citations

Cited by 25 Pith papers

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

  1. Flow Reasoning Models: Scaling Reasoning Through Iterative Self-Refinement

    cs.AI 2026-06 conditional novelty 7.0

    Flow models reach 99.2% Sudoku accuracy in 7 passes and 96.1% on out-of-distribution Sudoku-Extreme by selecting dynamically stable candidates and training with self-conditioning plus DPO to avoid failed outputs.

  2. TokenRatio: Principled Token-Level Preference Optimization via Ratio Matching

    cs.CL 2026-05 unverdicted novelty 7.0

    Introduces TBPO, which derives a Bregman-divergence density-ratio matching objective for token-level preference optimization that generalizes DPO while preserving the induced optimal policy.

  3. TokenRatio: Principled Token-Level Preference Optimization via Ratio Matching

    cs.CL 2026-05 unverdicted novelty 7.0

    TBPO posits a token-level Bradley-Terry model and derives a Bregman-divergence density-ratio matching loss that generalizes DPO while preserving token-level optimality.

  4. ResRL: Boosting LLM Reasoning via Negative Sample Projection Residual Reinforcement Learning

    cs.LG 2026-05 unverdicted novelty 7.0

    ResRL decouples shared semantics between positive and negative responses in LLM reinforcement learning via SVD-based projection residuals, outperforming baselines including NSR by up to 9.4% on math reasoning benchmarks.

  5. Step-level Denoising-time Diffusion Alignment with Multiple Objectives

    cs.LG 2026-04 unverdicted novelty 7.0

    MSDDA derives a closed-form optimal reverse denoising distribution for multi-objective diffusion alignment that is exactly equivalent to step-level RL fine-tuning with no approximation error.

  6. RAD-DPO: Robust Adaptive Denoising Direct Preference Optimization for Generative Retrieval in E-commerce

    cs.IR 2026-02 unverdicted novelty 7.0

    RAD-DPO adds token-level gradient detachment, similarity-based dynamic reward weighting, and a multi-label global contrastive objective to DPO for better handling of hierarchical Semantic IDs and noisy feedback in e-c...

  7. Test-Time Scaling via Error Localization

    cs.LG 2026-07 conditional novelty 6.0

    TTEL uses feedback-induced token probability drops to localize the first error in a failed reasoning trace and branch a new generation from that prefix, improving pass@k per token on coding and math benchmarks.

  8. Token-Level Off-Policy Learning for Faithful Generation Under Distribution Shift

    cs.CL 2026-07 conditional novelty 6.0

    A token-level correctness classifier trained with LoRA and then merged into the model boosts out-of-distribution factuality in summarization and translation.

  9. DyCo-RL: Dynamic Cross-Modal Coordination for Visual Reasoning

    cs.CV 2026-06 unverdicted novelty 6.0

    DyCo-RL improves four RLVR algorithms on seven visual and math reasoning benchmarks by assigning tokens visual or text roles via Fisher-Rao geodesic distance on attention and reweighting advantages by role-alignment score.

  10. TokenRatio: Principled Token-Level Preference Optimization via Ratio Matching

    cs.CL 2026-05 unverdicted novelty 6.0

    TBPO derives a token-level preference optimization objective from sequence-level pairwise data via Bregman divergence ratio matching that generalizes DPO and improves alignment quality.

  11. ResRL: Boosting LLM Reasoning via Negative Sample Projection Residual Reinforcement Learning

    cs.LG 2026-05 unverdicted novelty 6.0

    ResRL boosts LLM reasoning by modulating negative gradients with SVD-based projection residuals from negative samples, outperforming NSR by 9.4% Avg@16 on math benchmarks while preserving diversity across 12 tasks.

  12. Mobile GUI Agent Privacy Personalization with Trajectory Induced Preference Optimization

    cs.AI 2026-04 unverdicted novelty 6.0

    TIPO applies preference-intensity weighting and padding gating to stabilize preference optimization for privacy personalization in mobile GUI agents, yielding higher alignment and distinction metrics than prior methods.

  13. rePIRL: Learn PRM with Inverse RL for LLM Reasoning

    cs.LG 2026-02 unverdicted novelty 6.0

    rePIRL learns effective process reward models for LLM reasoning via a dual policy-PRM update process inspired by inverse RL, unifying online and offline methods with reported gains over prior approaches on math and co...

  14. 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.

  15. Distribution Preference Optimization: A Fine-grained Perspective for LLM Unlearning

    cs.LG 2025-10 conditional novelty 6.0

    DiPO is a distribution-level unlearning method that constructs preference distributions from the model's own high-confidence logits and achieves state-of-the-art forget quality on TOFU while preserving utility.

  16. FocusDPO: Dynamic Preference Optimization for Multi-Subject Personalized Image Generation via Adaptive Focus

    cs.CV 2025-09 conditional novelty 6.0

    FocusDPO adds dynamic spatial weighting to preference-based fine-tuning, improving subject fidelity and reducing attribute leakage in multi-subject personalized image generation.

  17. UNA: A Unified Supervised Framework for Efficient LLM Alignment Across Feedback Types

    cs.LG 2024-08 unverdicted novelty 6.0

    UNA unifies binary, pairwise, and score-based feedback for LLM alignment via a generalized implicit reward function shown optimal by the log sum inequality.

  18. 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...

  19. Enhancing Speech Large Language Models through Reinforced Behavior Alignment

    cs.CL 2025-08 unverdicted novelty 5.0

    Reinforced Behavior Alignment (RBA) uses self-synthesized data from a teacher LLM and reinforcement learning to close the instruction-following gap in SpeechLMs, outperforming distillation and reaching SOTA on spoken ...

  20. Difficulty-Based Preference Data Selection by DPO Implicit Reward Gap

    cs.CL 2025-08 unverdicted novelty 5.0

    Selecting preference pairs whose DPO implicit reward gap is small yields better LLM alignment than random or baseline selection while using only 10% of the data.

  21. From Fragments to Facts: A Curriculum-Driven DPO Approach for Generating Hindi News Veracity Explanations

    cs.CL 2025-07 unverdicted novelty 5.0

    A DPO framework augmented with curriculum learning and two new loss parameters generates veracity explanations for Hindi news using LLMs and PLMs.

  22. Analyzing and Improving Fine-grained Preference Optimization in Medical LVLMs

    cs.CV 2026-06 unverdicted novelty 4.0

    Proposes bidirectional token-wise KL regularizer and visual-contrastive grounding objective to create fine-grained on-policy preference pairs for medical LVLMs by minimally editing model outputs.

  23. rePIRL: Learn PRM with Inverse RL for LLM Reasoning

    cs.LG 2026-02 conditional novelty 4.0

    rePIRL learns token-level process rewards for LLM reasoning via a guided-cost-learning-style IRL objective, and shows these rewards improve reasoning policies on math/coding benchmarks.

  24. LLM Harms: A Taxonomy and Discussion

    cs.CY 2025-12 reject novelty 3.0

    Proposes a five-bucket taxonomy of LLM harms and calls for dynamic auditing, but the systematic review behind it is not reproducible and contains mismatched citations.

  25. LLM Harms: A Taxonomy and Discussion

    cs.CY 2025-12 unverdicted novelty 3.0

    This paper proposes a taxonomy of LLM harms in five categories and suggests mitigation strategies plus a dynamic auditing system for responsible development.