AlpsBench supplies 2500 real-dialogue sequences with verified memories to benchmark LLM extraction, updating, retrieval, and utilization of personalized information.
hub
Think-While-Generating: On-the-Fly Reasoning for Personalized Long-Form Generation.arXiv preprint arXiv:2512.06690.2025
19 Pith papers cite this work. Polarity classification is still indexing.
19
Pith papers citing it
hub tools
citation-role summary
baseline 1
citation-polarity summary
years
2026 19roles
baseline 1polarities
baseline 1representative citing papers
LLMs fail to capture embodied cognition and cultural variation in demonstrative use, unlike humans who show language-specific proximal-distal and perspective-taking patterns.
TriRec is a two-stage LLM-agent recommender that uses item self-promotion followed by platform-level sequential re-ranking to jointly optimize user utility, item exposure, and exposure fairness.
IntuRec anchors LLM latent reasoning for recommendation by deriving an intuition embedding from top-K candidates via self- and cross-attention to initialize more accurate trajectories.
Introduces ShopTrajQA long-context benchmark and an RLVR-trained tool-augmented agent that bypasses LLM context limits by external file storage and code-based retrieval for shopping trajectories.
PARL formulates personalized LLM evaluation as a learning problem that induces preference-aware rubrics from raw user histories via discriminative RL and self-validation.
PAD-Rec augments standard draft models with item-position and step-position embeddings plus learnable gates, delivering up to 3.1x wall-clock speedup and 5% average gain over strong speculative-decoding baselines on four datasets while largely preserving recommendation quality.
A cycle of solver, adversarial flawed-chain challenger, and feedback agent optimizes CoT prompts to raise accuracy and cut run-to-run answer variability within two to three cycles.
A score-ranking loss enables controllable summarization by aligning outputs to evaluation scores, matching SOTA performance with dimension-specific control on LLaMA, Qwen, and Mistral.
Agent-GWO uses collaborative grey-wolf-inspired agents to jointly optimize LLM prompts and decoding settings, yielding higher accuracy and stability than prior single-agent prompt optimization methods on math and hybrid reasoning benchmarks.
FACT-E uses controlled perturbations as an instrumental signal to measure intra-chain faithfulness in CoT reasoning and combines it with answer consistency to select trustworthy trajectories.
PUMA applies the Free Energy Principle to maintain beliefs over latent user states and select actions by minimizing expected free energy in multi-turn personalized dialogues.
CAP-CoT improves LLM reasoning accuracy and stability by iteratively refining solver prompts via contrast with adversarially generated flawed reasoning chains.
A hybrid graph-based training-free framework for LLM context compression matches strong baselines and shows larger gains on long-document benchmarks.
A reference-free proxy scoring framework combined with GIRB calibration produces better-aligned evaluation metrics for summarization and outperforms baselines across seven datasets.
AGSC combines NLI neutral probabilities for adaptive granularity with GMM semantic clustering to improve uncertainty quantification in long-text LLM generation, claiming SOTA factuality correlation and 60% faster inference.
LLMs show strong exam performance on medical tasks but exhibit a clear gap in accuracy on authentic clinical decision-making as measured by the new MR-Bench benchmark and unified evaluations.
InSemRAG combines dynamic intent-aware hybrid retrieval and semantics-preserving chunk repair in an iterative loop, yielding 2.65 F1 gain on HotPotQA and 1.5 accuracy gain on FEVER with 4.32x lower latency than Multi-Hop RAG via SLMs.
A small language model resolves semantic risks and conflicts in prompts via multi-perspective consistency checks, yielding a 2.5-point gain in LLM reasoning performance at $0.02 cost.
citing papers explorer
-
AlpsBench: An LLM Personalization Benchmark for Real-Dialogue Memorization and Preference Alignment
AlpsBench supplies 2500 real-dialogue sequences with verified memories to benchmark LLM extraction, updating, retrieval, and utilization of personalized information.
-
Do LLMs Capture Embodied Cognition and Cultural Variation? Cross-Linguistic Evidence from Demonstratives
LLMs fail to capture embodied cognition and cultural variation in demonstrative use, unlike humans who show language-specific proximal-distal and perspective-taking patterns.
-
Breaking User-Centric Agency: A Tri-Party Framework for Agent-Based Recommendation
TriRec is a two-stage LLM-agent recommender that uses item self-promotion followed by platform-level sequential re-ranking to jointly optimize user utility, item exposure, and exposure fairness.
-
Intuition-Guided Latent Reasoning for LLM-Based Recommendation
IntuRec anchors LLM latent reasoning for recommendation by deriving an intuition embedding from top-K candidates via self- and cross-attention to initialize more accurate trajectories.
-
Customer-Agent: Overcoming Context Limitations in Ultra-Long Shopping Trajectories via Tool-Augmented Agents and RLVR
Introduces ShopTrajQA long-context benchmark and an RLVR-trained tool-augmented agent that bypasses LLM context limits by external file storage and code-based retrieval for shopping trajectories.
-
Preference-Aware Rubric Learning for Personalized Evaluation
PARL formulates personalized LLM evaluation as a learning problem that induces preference-aware rubrics from raw user histories via discriminative RL and self-validation.
-
Position-Aware Drafting for Inference Acceleration in LLM-Based Generative List-Wise Recommendation
PAD-Rec augments standard draft models with item-position and step-position embeddings plus learnable gates, delivering up to 3.1x wall-clock speedup and 5% average gain over strong speculative-decoding baselines on four datasets while largely preserving recommendation quality.
-
CAP-CoT: Cycle Adversarial Prompt for Improving Chain of Thoughts in LLM Reasoning
A cycle of solver, adversarial flawed-chain challenger, and feedback agent optimizes CoT prompts to raise accuracy and cut run-to-run answer variability within two to three cycles.
-
Learning to Control Summaries with Score Ranking
A score-ranking loss enables controllable summarization by aligning outputs to evaluation scores, matching SOTA performance with dimension-specific control on LLaMA, Qwen, and Mistral.
-
Agent-GWO: Collaborative Agents for Dynamic Prompt Optimization in Large Language Models
Agent-GWO uses collaborative grey-wolf-inspired agents to jointly optimize LLM prompts and decoding settings, yielding higher accuracy and stability than prior single-agent prompt optimization methods on math and hybrid reasoning benchmarks.
-
FACT-E: Causality-Inspired Evaluation for Trustworthy Chain-of-Thought Reasoning
FACT-E uses controlled perturbations as an instrumental signal to measure intra-chain faithfulness in CoT reasoning and combines it with answer consistency to select trustworthy trajectories.
-
Know You Before You Speak: User-State Modeling for LLM Personalization in Multi-Turn Conversation
PUMA applies the Free Energy Principle to maintain beliefs over latent user states and select actions by minimizing expected free energy in multi-turn personalized dialogues.
-
Partial exploiters sustain cooperation: the hump-shaped strategy stably coexists with unconditional cooperators
CAP-CoT improves LLM reasoning accuracy and stability by iteratively refining solver prompts via contrast with adversarially generated flawed reasoning chains.
-
From Similarity to Structure: Training-free LLM Context Compression with Hybrid Graph Priors
A hybrid graph-based training-free framework for LLM context compression matches strong baselines and shows larger gains on long-document benchmarks.
-
Calibrating Model-Based Evaluation Metrics for Summarization
A reference-free proxy scoring framework combined with GIRB calibration produces better-aligned evaluation metrics for summarization and outperforms baselines across seven datasets.
-
AGSC: Adaptive Granularity and Semantic Clustering for Uncertainty Quantification in Long-text Generation
AGSC combines NLI neutral probabilities for adaptive granularity with GMM semantic clustering to improve uncertainty quantification in long-text LLM generation, claiming SOTA factuality correlation and 60% faster inference.
-
Medical Reasoning with Large Language Models: A Survey and MR-Bench
LLMs show strong exam performance on medical tasks but exhibit a clear gap in accuracy on authentic clinical decision-making as measured by the new MR-Bench benchmark and unified evaluations.
-
Efficient RAG with Intent-Aware Retrieval and Semantics-Preserving Chunking
InSemRAG combines dynamic intent-aware hybrid retrieval and semantics-preserving chunk repair in an iterative loop, yielding 2.65 F1 gain on HotPotQA and 1.5 accuracy gain on FEVER with 4.32x lower latency than Multi-Hop RAG via SLMs.
-
Small Language Model Helps Resolve Semantic Ambiguity of LLM Prompt
A small language model resolves semantic risks and conflicts in prompts via multi-perspective consistency checks, yielding a 2.5-point gain in LLM reasoning performance at $0.02 cost.