Post-hoc Alignment of LLM-judges to Human Judgment Distribution

• URL: http://arxiv.org/abs/2609.01073v1
• Date: Tue, 01 Sep 2026 11:03:56 GMT
• Title: Post-hoc Alignment of LLM-judges to Human Judgment Distribution
• Abstract summary: We show that NAPHA consistently improves soft-labels prediction across base LLM models and datasets.<n>We also show via oracle experiments that improving entropy class prediction can substantially enhance NAPHA's practical effectiveness.
• Score: 3.0607199893136006
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: The LLM-as-a-judge (LLMaJ) framework offers a cost-effective and reproducible solution for automatic evaluation. However, current evaluation practices typically compare LLMaJ judgments against aggregated ground-truth labels, overlooking the valuable information contained in Human Label Variation (HLV). Inspired by an increasing line of work that proposes to leverage HLV, we systematically study LLMaJ performance on predicting both a single, aggregated ground truth hard-label and unaggregated soft-labels that represent Human Judgment Distributions (HJD). Our results across five diverse datasets reveal that while LLMs achieve near human-level performance at hard-label prediction on most tasks, they exhibit poor performance when predicting soft-labels. To address this limitation, we propose NAPHA (eNtropy-Aware Post-Hoc Alignment), a simple yet effective lightweight post-hoc alignment method that matches the LLM distribution to the HJD by first assigning an instance to a discrete entropy class and then routing it to specialized, trained alignment models. We find that NAPHA consistently improves soft-labels prediction across base LLM models and datasets, with particularly strong gains on high-entropy instances where capturing diverse human perspectives is most critical. We also show via oracle experiments that improving entropy class prediction can substantially enhance NAPHA's practical effectiveness.

Related papers

• Rethinking the Evaluation and Optimization of LLM-Based Social Simulation [56.18577997572273]
  We use the subjectivity coefficient to analyze how accuracy-based evaluation and hard-label training fail as subjectivity grows.<n>We construct SUBJSIM, a benchmark of 19,300 contexts covering 193 annotators and 100 subjective questions.<n>Results on SUBJSIM demonstrate the advantages of our method.
  arXiv  Detail & Related papers  (2026-08-20T06:26:03Z)
• DAJ: Data-Reweighted LLM Judge for Test-Time Scaling in Code Generation [30.131052926559956]
  We propose DAJ, a reasoning-based LLM judge trained with rewards under a bi-level data-reweighted learning framework.<n>Our approach automatically emphasizes hard problems, in-distribution samples, and trajectory-aligned data, without relying on hand-crafted verifiables.
  arXiv  Detail & Related papers  (2026-01-29T19:04:24Z)
• LLM Performance Predictors: Learning When to Escalate in Hybrid Human-AI Moderation Systems [5.7001352660257005]
  We propose a framework for supervised uncertainty quantification in content moderation systems.<n>We show that our method enables cost-aware selective classification in real-world human-AI.<n>This work establishes a principled framework for uncertainty-aware, scalable and responsible human-AI moderation.
  arXiv  Detail & Related papers  (2026-01-11T17:46:49Z)
• GEM: Generative Entropy-Guided Preference Modeling for Few-shot Alignment of LLMs [5.1816417820270075]
  In this paper, we propose a generative entropy-guided preference modeling approach named GEM for large language models (LLMs)<n>Instead of training a discriminative reward model on preference data, we directly train the LLM to internalize a closed-loop optimization architecture.<n> Experiments on general benchmarks and domain-specific tasks demonstrate that our GEM achieves significant improvements with few-shot preference data.
  arXiv  Detail & Related papers  (2025-11-17T06:04:47Z)
• Beyond the Surface: Enhancing LLM-as-a-Judge Alignment with Human via Internal Representations [15.542741121573203]
  LAGER is a framework for improving the alignment of "LLM-as-a-Judge" evaluations with human scores by leveraging internal representations.<n>We evaluate our method on the standard alignment benchmarks Flask, HelpSteer, and BIGGen using Spearman correlation, and find that LAGER achieves improvements of up to 7.5% over the best baseline.
  arXiv  Detail & Related papers  (2025-08-05T15:18:36Z)
• Mitigating Forgetting in LLM Fine-Tuning via Low-Perplexity Token Learning [65.23593936798662]
  We show that fine-tuning with LLM-generated data improves target task performance and reduces non-target task degradation.<n>This is the first work to provide an empirical explanation based on token perplexity reduction to mitigate catastrophic forgetting in LLMs after fine-tuning.
  arXiv  Detail & Related papers  (2025-01-24T08:18:56Z)
• Reward-Augmented Data Enhances Direct Preference Alignment of LLMs [63.32585910975191]
  We introduce reward-conditioned Large Language Models (LLMs) that learn from the entire spectrum of response quality within the dataset.<n>We show that our approach consistently boosts DPO by a considerable margin.<n>Our method not only maximizes the utility of preference data but also mitigates the issue of unlearning, demonstrating its broad effectiveness beyond mere data expansion.
  arXiv  Detail & Related papers  (2024-10-10T16:01:51Z)
• Uncertainty Aware Learning for Language Model Alignment [97.36361196793929]
  We propose uncertainty-aware learning (UAL) to improve the model alignment of different task scenarios. We implement UAL in a simple fashion -- adaptively setting the label smoothing value of training according to the uncertainty of individual samples. Experiments on widely used benchmarks demonstrate that our UAL significantly and consistently outperforms standard supervised fine-tuning.
  arXiv  Detail & Related papers  (2024-06-07T11:37:45Z)
• Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment [72.99676237703099]
  We propose a new framework that boosts the alignment of large language models with human preferences.<n>Our key idea is leveraging the human prior knowledge within the small (seed) data.<n>We introduce a noise-aware preference learning algorithm to mitigate the risk of low quality within generated preference data.
  arXiv  Detail & Related papers  (2024-06-06T18:01:02Z)
• A Channel-ensemble Approach: Unbiased and Low-variance Pseudo-labels is Critical for Semi-supervised Classification [61.473485511491795]
  Semi-supervised learning (SSL) is a practical challenge in computer vision. Pseudo-label (PL) methods, e.g., FixMatch and FreeMatch, obtain the State Of The Art (SOTA) performances in SSL. We propose a lightweight channel-based ensemble method to consolidate multiple inferior PLs into the theoretically guaranteed unbiased and low-variance one.
  arXiv  Detail & Related papers  (2024-03-27T09:49:37Z)
• From Quantity to Quality: Boosting LLM Performance with Self-Guided Data Selection for Instruction Tuning [52.257422715393574]
  We introduce a self-guided methodology for Large Language Models (LLMs) to autonomously discern and select cherry samples from open-source datasets. Our key innovation, the Instruction-Following Difficulty (IFD) metric, emerges as a pivotal metric to identify discrepancies between a model's expected responses and its intrinsic generation capability.
  arXiv  Detail & Related papers  (2023-08-23T09:45:29Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.