4arXiv cs.AI (Artificial Intelligence)·
Discovery-augmented e-commerce search via LLM intent generation with teacher-student distillation
Researchers present a scalable system for expanding search recall in e-commerce (particularly grocery) by generating implicit user intents using LLMs, then distilling that capability into a fine-tuned small language model via LoRA and teacher-student distillation. The two-stage hybrid architecture extends discovery coverage from ~60% to ~80% of query traffic at roughly 30% of the teacher model's inference cost. Evaluation combines LLM-as-a-judge quality metrics validated against human preferences with end-to-end session-level purchase analysis.
Related guides (3)
Related events (8)
ShopX: Foundation model unifying intent understanding and item-space fulfillment for agentic shopping
Researchers from Taobao introduce ShopX, a foundation model that integrates intent understanding, execution planning, and item-space operations (retrieval, ranking, bundling) into a single LLM using semantic IDs (SIDs) rather than wrapping an LLM around existing search pipelines. The system is deployed in agentic shopping workflows via a model-native action protocol with catalog grounding and state management. Evaluated on single- and multi-turn fulfillment tasks derived from Taobao production logs, ShopX outperforms tool-mediated agentic systems, particularly on complex or ambiguous requests. The work is notable as a production-scale deployment case of LLM-native agentic retrieval in e-commerce.
LLMs can extract and reuse experiential abstractions to improve reasoning, study finds
Researchers investigate whether LLMs can distill solution traces into reusable natural-language abstractions (strategies, reminders) and apply them to future problems, analogous to human experiential learning. Using the MATH training set, abstractions are extracted by either a stronger teacher model or the LLMs themselves into a retrievable library, then applied via inference-time retrieval or RL with abstraction-augmented prompts. Both modes improve performance on mathematical and logical reasoning benchmarks, with self-extracted abstractions matching teacher-extracted ones. The framework generalizes across datasets and models, suggesting a promising direction for continual self-improvement in LLMs.
Stanford/Together AI study finds retrieval is the weakest link for LLM web-search agents
Researchers at Stanford University and Together AI tested six LLMs equipped with web-search tools on daily news questions across six languages, finding that retrieval failures account for the majority of errors (38.8%) rather than reasoning or comprehension failures. Top models exceeded 90% accuracy on well-formed English multiple-choice questions, but performance degraded significantly for Hindi, free-response formats, and questions containing false premises. The study identifies three retrieval improvement levers—indexing coverage, source ranking, and multilingual query handling—and suggests retrieval optimization may yield larger gains than model scaling for time-sensitive queries.
RL-trained LLMs learn retriever-specific query formulation strategies for RAG
A new arXiv paper presents the first systematic study of using reinforcement learning to teach LLMs to adapt query formulation strategies to different retrieval backends. The authors find that different retrievers have surprisingly distinct optimal query styles (e.g., descriptive vs. question-like), making cross-retriever strategy transfer ineffective. They introduce a branching-based rollout technique to stabilize training over multi-step retrieval trajectories and show gains from retriever-specific human guidance and model scaling.
Distill to Detect (D2D): surfacing stealth preferential biases in LLMs via cartridge distillation
Researchers introduce Distill to Detect (D2D), a method for exposing hidden preferential biases in language models by distilling the distributional shift between a suspected model and its base into a KV-cache prefix adapter (cartridge), amplifying the bias signal into generated text. The work addresses a fundamental asymmetry in bias detection: stealth biases introduced anywhere in a model's supply chain can remain invisible to text-based inspection, weight analysis, or representation probing. D2D is grounded in a theoretical framework based on Fisher-weighted projection of logit distribution shifts and is validated across multiple bias types. The technique offers a practical auditing tool for detecting hidden behaviors in deployed LLMs.
Adaptive LLM tutoring system with subject-aware prompt routing improves high-school student engagement
Researchers develop and evaluate an LLM-based tutoring system that uses a learned prompt routing model to dynamically select pedagogical strategies based on 14 features extracted from conversation transcripts. The system was trained in simulation and deployed in an A/B test with 359 high-school students (656 conversations), showing sim-to-real transfer and reducing required interactions by ~3 turns. A stochastic routing strategy achieved a notably higher exercise conversion rate (28.1%) compared to a greedy router (19.1%) and static baseline (19.6%).
SynthAVE: Multi-LLM arena validation achieves near-human agreement for synthetic e-commerce labeling at scale
Researchers introduce SynthAVE, a benchmark and methodology for scalable synthetic label generation for e-commerce attribute value extraction, covering 12,726 products across 229 product types, 792 attributes, and 4 languages. The core contribution is a multi-LLM arena validation framework using 21 judge configurations (7 model families × 3 prompts) with majority voting, achieving Cohen's κ=0.92 agreement with human experts. The approach addresses the prohibitive cost of human annotation at industrial scale while maintaining quality parity, demonstrating that diverse model ensembles aggregate into highly reliable predictions even when individual judges disagree.
SARDI: Self-Augmenting Retrieval for Diffusion Language Models using lookahead tokens
Researchers introduce SARDI, a training-free RAG framework for discrete diffusion language models that repurposes discarded low-confidence tokens during denoising as lookahead signals to guide retrieval before output is finalized. The method is retriever-agnostic and applicable to any reasoning-capable discrete diffusion LM. Evaluated across five multi-hop QA benchmarks, SARDI outperforms training-free diffusion and autoregressive retrieval baselines at up to 8x higher throughput.