LIVE · refreshes every 20 min
updated Nov 14, 07:23 UTC
110010.art(ificial intelligence)
AI products, research, and launches — clustered and ranked, not just re-blogged.
110010.art / archive

Research

Nov 10
4d ago How many examples, and which ones? A study of in-context learning for natural language inference arXiv cs.CL → story
4d ago I think, therefore I attack: evaluating thought injection attacks on reasoning agents arXiv cs.AI → story
4d ago Implicit bias in state space models and linear autoregressive training arXiv cs.LG → story
4d ago Is Recursive Self-Improvement Safe? arXiv cs.AI → story
4d ago IWIBench evaluates a large editor model’s ability to modify another model’s weights through intra-weight injection. arXiv cs.AI → story
4d ago Jacobian-Causal Memory Control for persistent memory in language model systems arXiv cs.AI → story
4d ago Jais 3 releases six Arabic-centric agentic language models from 0.9B to 375B parameters; capable of planning multi-step tasks, tool use, web search, and document editing in Arabic arXiv cs.CL → story
4d ago Key institutional considerations for designing effective independent verification organizations arXiv cs.AI → story
4d ago Learning Across heterogeneous feature spaces with group robustness arXiv cs.LG → story
4d ago Learning of Tree Automata applied to Neural Language Acceptors arXiv cs.CL → story
4d ago Learning whom to trust about the unseen: source credibility and reliability transfer in language models arXiv cs.AI → story
4d ago LLM-based feature extraction for multi-criteria inventory classification arXiv cs.AI → story
4d ago LLMs annotate veteran crisis severity; study compares 15 open-source and 10 proprietary models to eight veteran SMEs using 149 posts arXiv cs.AI → story
4d ago LongCapHarness: capability-guided auto-harnessing for long-context reasoning arXiv cs.CL → story
4d ago LoRazor: trimming adapter execution for efficient multi-LoRA serving arXiv cs.AI → story
4d ago MAGE: towards generalizable multi-timescale EEG representations arXiv cs.LG → story
4d ago Measuring prompt injection guards on legal contract review attacks; false positive rates reach 92.39% on contract data, and guards underperform on evaluation metrics arXiv cs.AI → story
4d ago MedToolFault and MedGuard: evaluating and improving clinical agent reliability arXiv cs.CL → story
4d ago Memory-aware multi-agent architecture with LLM-based orchestration for conversational recommendation arXiv cs.AI → story
4d ago META-IR: Meta-Introspective Reasoning for Adaptive Safety Alignment in LLMs arXiv cs.AI → story
4d ago MIT-led team receives funding to develop next-generation alternatives to modern weed killers MIT News (AI) → story
4d ago NormGym: evaluating normative competence through social interaction arXiv cs.AI → story
4d ago PACE: Keeping the right tokens in the KV cache is not enough for long-horizon LLM agents arXiv cs.LG → story
4d ago Pain states in language models can be injected and reported; three open-weight models tested with OWMI reveal amplification of pain directions drives destructive actions. arXiv cs.CL → story
4d ago Pdb or Not to Be: comparing debugging interfaces with Pass$_{X}$@$k$ arXiv cs.AI → story
4d ago Persistent behavioral learning for AI agents with Forge Decision Context captures human corrections and self-assessed tool-call failures into imperative knowledge; arXiv:2611.08644v1 arXiv cs.AI → story
4d ago Process then retrieve: investigating the origins of the modality gap in vision-language models arXiv cs.LG → story
4d ago Projected neural additive models as universal approximators arXiv cs.LG → story
4d ago Prompt-space recovery of knowledge edits in language models enables recovering pre-edit knowledge through input-space interventions without modifying model weights using Memento arXiv cs.CL → story
4d ago ProofDiff analyzes auditable, argument-level differences in AI-generated mathematical proofs arXiv cs.AI → story
4d ago RAD-TFM: Robust and Domain-Adapted Tabular Foundation Models arXiv cs.LG → story
4d ago Rethinking agent trajectory compression: a workflow-aware perspective arXiv cs.AI → story
4d ago Rethinking how to leverage cheap physics for data-efficient neural operators arXiv cs.LG → story
4d ago Rethinking supervision targets for multimodal LLM unlearning: less private information, reduced distribution shift arXiv cs.CL → story
4d ago Retrofitting unbounded context onto stock language models; the paper shows two stock models can retain and query across all prior context by manipulating geometry through Rotary attention arXiv cs.CL → story
4d ago RewardBench: evaluating reward models for deep research reports arXiv cs.AI → story
4d ago ScreenProof detects UI faults on screenshots without a reference, identifying issues like off-screen labels, touching letters, broken lines, and mismatched button corner styles arXiv cs.AI → story
4d ago Selected-Risk Calibration for feasibility at returned decisions arXiv cs.LG → story
4d ago Semiprime factorization with discrete diffusion and beam search achieves 68-bit factoring with learning-based approach; surpasses 56-bit state of the art arXiv cs.LG → story
4d ago Solver-admitted metamorphic testing of learned network optimization surrogates; arXiv study examines equivalence-encoding differences, numerical error, and constraint checks for audit eligibility arXiv cs.AI → story

Older items: monthly archive →