Skip to main content
Search

University of Illinois Chicago

10 related articles

Subscribe to updates
  1. arXiv

    Agent memory stops only looking back: Pro-Memory splits future intentions from the moment to act

    Most memory systems for large language model agents only store and retrieve past information, while proactive agents also need to hold future intentions and realize them when the right moment arrives. Pro-Memory separates keeping a future intention from deciding when and how to realize it: pending objectives stay linked to their supporting evidence, new observations are checked against the original goal and activation conditions, and contributions are coordinated with the agent's ongoing actions. On two benchmarks covering delayed-intention tasks and open-ended proactive assistance, proactive task performance improves consistently over representative memory approaches across multiple LLM backbones while unnecessary interventions are limited.

  2. arXiv

    Instruction tuning on 0.5% of the data: QSGS cuts redundancy with a graph cut and matches full-data Mistral-7B

    For targeted instruction-tuning data selection, QSGS organizes candidate examples by gradient similarity into a graph and applies a submodular graph cut that pulls the subset toward the target task while penalizing internal redundancy, then adds gradient-guided synthesis to fill directions the pool cannot cover. On MMLU, BBH and TydiQA across Llama-2, Llama-3 and Mistral, QSGS at a 0.5% budget (1,350 of 270,000 examples) matches or exceeds full-data performance on Mistral-7B and beats the strongest baseline on every benchmark.

  3. arXiv

    Stance prompts stop being rewritten whole: six modules tuned one at a time, errors routed by label, four of five backbones gain

    Stance-detection prompt optimization now splits the prompt into six functional modules refined one at a time, with misclassification feedback routed by gold label, and on SemEval-2016 four of five open-weight backbones reach higher mean Macro-F1 than pooling all errors together.

  4. arXiv

    SWR self-supervises training environments from 500 scientific software workflows, lifting Qwen3.8-27B on Terminal-Bench 2 from 47.94% to 53.56%

    The authors introduce software-in-the-loop reconstruction (SWR), a self-supervised framework that draws reference outputs and verification targets from existing software workflows, instantiated with 500 workflows across 46 software families and six domains, where Qwen3.8-Max solves 838 tasks over three attempts and yields 1,422 verified trajectories oversampled into 3,000 reconstruction-only training examples, and supervised fine-tuning raises Qwen3.8-27B mean Terminal-Bench 2 performance from 47.94% to 53.56% while achieving the highest mean among four matched-token corpus controls on all four reported evaluations.

  5. arXiv

    Fisher-guided submodular data selection lets 1B selected tokens beat 10B replay tokens on both adaptation and forgetting in medical continual pre-training

    The work shows that loss-based data selection drives the post-continual-pre-training Fisher diagonal downward on exactly the high-Fisher coordinates the pretrained model had committed to while leaving low-Fisher coordinates largely untouched, and proposes a Fisher-aware selector that decomposes each candidate's gradient into an anchor component and a frontier component and aggregates them with a log-determinant submodular objective optimized in a single streaming pass; on TinyLlama-1.1B and Llama-3.1-8B continual pre-training over medical data it improves target-domain quality while bounding forgetting on held-out pretraining benchmarks, and 1B selected tokens outperform a replay strategy trained with 10B tokens on both adaptation and forgetting, a roughly 10x token-efficiency advantage.

  6. arXiv

    GRAFT lets one backbone absorb five vision foundation models in sequence, with each new capability costing a single distillation

    GRAFT introduces a continual multi-teacher knowledge distillation framework in which the previously distilled model acts as a teacher to preserve learned capabilities while the current student jointly learns from that model and the incoming teacher, and it adds Teacher Specific Readout Tokens plus a Geometry Agnostic Relational Loss to reconcile the incompatible representation geometries of heterogeneous teachers, yielding a single continually extensible backbone that unifies image understanding, 2D dense prediction, 3D human pose estimation, 3D vision, and vision-language, with each new capability acquired at the cost of a single distillation rather than a full re-distillation.

  7. arXiv

    Latent-MOPD lets one student absorb both the predictions and hidden states of multiple specialists, beating single-channel baselines on all nine math, code, and logic benchmarks

    The work introduces Latent-MOPD, a multi-teacher on-policy distillation method for LLMs that, to the authors' knowledge, is the first to integrate specialists at the representation level: it selects late-layer hidden-state targets according to the teacher-student relationship, bridges unequal hidden widths with a shared projection, groups updates by domain, and gradually shifts each teacher's supervision from hidden states to token predictions; in the same-family setting it outperforms token-only, representation-only, and uniform-averaging baselines on all nine benchmarks across math, code, and logic, surpasses the per-benchmark best teacher on a majority of them at equal parameter count, and with larger cross-family teachers it outperforms both single-channel baselines on all benchmarks.

  8. A 45-member modified Delphi consensus: screen for eating disorders before GLP-1 receptor agonist prescribing, and generally avoid these agents in active anorexia and bulimia nervosa

    Using a three-round modified Delphi method, a 45-member panel of clinicians and researchers in diabetes, obesity and/or eating disorders plus people with lived experience reached an average 91.7% consensus on recommendations for GLP-1 receptor agonist use in eating disorder contexts, covering pre-treatment screening, risk-stratified monitoring, standardized educational materials and concurrent evidence-based psychotherapy, stating that these agents should generally be avoided in active eating disorders, particularly anorexia nervosa, atypical anorexia nervosa and bulimia nervosa with marked dietary restraint and/or significant weight suppression, and setting out research priorities in epidemiology, regulation and binge eating disorder trials.

  9. arXiv

    TraceDance auto-builds 107 behavior benchmarks from 252,557 real deployment traces, and nine frontier models pass only 26.7% on average

    TraceDance is an agent system that constructs targeted benchmarks from real deployment traces for user-specified undesirable behaviors, using Anchor-and-Confirm (programmable retrieval plus candidate-level confirmation by a Flash LLM) and an Anchor Synthesis Loop to generate behavior specifications, and using decision-point continuation so an evaluated LLM produces its next turn at a recorded decision point and is graded by a behavior-specific rubric; experiments on 252,557 sessions from Claude Code and OpenClaw produce 107 benchmarks with 4,125 instances, fulfilling 95.3% of build-target requests, with both human annotators confirming the requested behavior in 84% of sampled instances, while nine frontier LLMs achieve a mean pass rate of only 26.7%.

  10. arXiv

    Half-Moon Cookie lets a sender run a one-shot private approximate blocklist check while receivers confirm it with an implicit check as fast as 0.19 s, resisting TOCTOU attacks

    The work introduces Half-Moon Cookie, a three-party framework in which a sending client performs a privacy-preserving approximate (metric-space) check of an item against a server's proprietary blocklist; on success the server stores a hiding and binding token in an allowlist, and a receiving client can later confirm via a much faster implicit check that the item still passes, mitigating TOCTOU attacks without revealing client inputs or the blocklist; the authors instantiate it for Hamming-distance blocklists and apply it to similarity-based malware detection, showing reusable garbled circuits cut embedding communication by over two orders of magnitude and that the implicit check needs only 7.2e-3 MB and 0.19 s on a 100 kB input.

You’re all caught up