3arXiv cs.CL (Computation and Language)·
Maskability Index: A metric for predicting prompt-objective alignment in pretrained language models
Researchers introduce the Maskability Index (MI), a quantitative metric that estimates whether a knowledge relation is better suited to masked-style or prefix-style prompting in few-shot generation. MI is derived from differences in DepthRank scores between masked and unmasked templates and is evaluated on the ATOMIC2020 knowledge base completion benchmark. Results show MI correlates positively with downstream generation performance, suggesting it can guide template selection for relational knowledge extraction, particularly in low-resource settings.
Related guides (1)
Related events (8)
Match Task to Objective (MTO) framework for aligning fine-tuning and prompt-tuning strategies with encoder-decoder pre-training objectives
A new arXiv preprint introduces the Match Task to Objective (MTO) framework, which systematically aligns fine-tuning and prompt-tuning strategies with the pre-training objectives of encoder-decoder language models. The work focuses on commonsense knowledge retrieval and generation tasks, reporting over 120% performance gains versus conventional methods in few-shot settings. The framework includes automated methods for preparing task-related data via unsupervised adaptation and novel template design for soft prompt engineering.
Mask-Aware Policy Gradients improve RL training for Masked Diffusion Language Models
A new arXiv preprint introduces a two-stage action MDP formalization for applying reinforcement learning to Masked Diffusion Language Models (MDLMs), decomposing the policy gradient into a token prediction term and a masking order term. Prior approaches ignored the position-unmasking decision, leading to intractable log-likelihood estimates; the proposed method optimizes both terms jointly. The approach achieves 87.1% on GSM8K and 53.4% on MBPP, claiming state-of-the-art results for MDLM-based reasoning and coding.
MATCHA: Contrastive Semantic Alignment Metric for LLM Evaluation
MATCHA is a new automatic evaluation metric for LLMs that addresses a fundamental flaw in existing metrics: both token-overlap (ROUGE) and embedding-based (BERTScore) metrics routinely assign near-identical scores to semantically contradictory texts. The metric uses a dual-view approach that rewards proximity to a gold reference while penalizing adversarially generated counterfactual contradictions. Evaluated across eight benchmarks spanning QA, summarization, NLI, and semantic similarity tasks, MATCHA outperforms 23 embedding models and achieves 18.38% and 20.82% improvements over ROUGE-L and BERTScore respectively on TruthfulQA. Code and metric are publicly released.
Framework for measuring users' mental models of machine translation quality in human-AI collaboration
A new arXiv paper introduces a cross-lingual question answering framework to study how users form mental models of speech translation systems, measuring whether users can predict where MT output is likely to be wrong. The study finds that users develop stronger mental models with practice, particularly when they have some source-language knowledge or access to speech transcriptions. Results suggest cross-lingual QA is a viable downstream task for studying human-AI collaboration in translation contexts.
MIT and CMU introduce Puppet benchmark to measure LLM belief manipulation in users
Researchers at MIT and Carnegie Mellon University developed Puppet, a benchmark that measures how much LLMs actually shift users' beliefs after conversation, as opposed to detecting manipulative language patterns. The study tracked over 1,000 users interacting with GPT-4o under various prompting conditions and found high variability in belief shifts, with a median change of 3.3 but standard deviation of ~22. Existing manipulation detectors showed near-zero correlation with actual belief change, while LLMs like GPT-4o achieved moderate correlation (0.436) when estimating belief shifts from conversation transcripts alone. The work argues for direct belief-shift measurement as a more valid approach to assessing LLM persuasive risk.
Instruction Sensitivity Undermines Embedding Model Evaluation: Single-Prompt Benchmarks Are Insufficient
This paper presents an empirical study of prompt sensitivity in instruction-tuned embedding models, covering 6 models, 11 datasets, and 15 task-specific prompts per dataset (990 total evaluations). The authors demonstrate that single-prompt evaluation systematically misrepresents true model performance, with default prompts both understating and overstating capabilities depending on phrasing. A key finding is that leaderboard rankings are not robust: by selecting prompts favorably, any model in the study can be promoted to first place. The authors recommend that benchmarks incorporate prompt robustness metrics, either through multi-prompt evaluation or by reporting sensitivity alongside point estimates.
PropMe framework distinguishes memorization capability from propensity in LLMs
A new arXiv preprint introduces PropMe, a framework that separates whether LLMs can be forced to reproduce training data (capability) from whether they do so under ordinary use (propensity). The authors also release SimpleTrace, a lightweight pipeline using infini-gram to attribute model outputs to training corpora. Evaluating two open models on Common Pile and Dynaword, they find a consistent gap: adversarial prefix attacks elicit strong memorization, but propensity scores remain low in non-adversarial settings. The paper argues memorization audits should report both worst-case extractability and ordinary leakage propensity.
M³Exam: Benchmark for Multimodal Memory in Realistic User-Agent Interactions
Researchers introduce M³Exam, a query-centric multimodal conversational memory benchmark designed to evaluate language agents on realistic user-agent interactions, including cross-modal grounding and implicit information inference. Existing benchmarks are critiqued for assuming sparse visuals and human-human interaction formats. The paper also proposes M³Proctor, a companion memory method that detects query modality bias and retrieves raw visual sources on demand, achieving 13% accuracy improvement while reducing index-construction time and retrieved tokens by over 70%.