大模型基准论文趋势 2023–2026

主要结论

• 大模型基准论文从 2023 年的 2 篇增长到 2026 年的 307 篇，三年间扩大到约 154 倍。
• NeurIPS 是发大模型基准论文最多的会议，2023–2026 年共 256 篇，占四个会议总量的 47%。
• 交叉方向里涨得最快的是AI 智能体：同时属于两个主题的论文从 2023 年的 1 篇增至 2026 年的 116 篇。
• 在「评测与基准」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的模型即裁判是 27 倍，大模型基准是 154 倍。

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

• NeurIPS 2026A Free Lunch in LLM Compression: Revisiting Retraining after Pruning
• ICML 2026Controlled LLM Training on Spectral Sphere
• ICLR 2026Hey, That's My Model! Introducing Chain & Hash, An LLM Fingerprinting Technique
• CVPR 20263DrawAgent Teaching LLM to Draw in 3D with Early Contrastive
• NeurIPS 2025ACCO: Accumulate While You Communicate for Communication-Overlapped Sharded LLM Training
• ICML 2025Critical Tokens Matter: Token-Level Contrastive Estimation Enhances LLM’s Reasoning Capability
• ICLR 2025Catastrophic Failure of LLM Unlearning via Quantization
• CVPR 2025Period-LLM: Extending the Periodic Capability of Multimodal Large Language Model
• NeurIPS 2024Large Language Models' Expert-level Global History Knowledge Benchmark (HiST-LLM)
• ICML 2024Characterizing ResNet's Universal Approximation Capability
• ICLR 2024Faithful Explanations of Black-box NLP Models Using LLM-generated Counterfactuals
• CVPR 2024Koala: Key Frame-Conditioned Long Video-LLM
• NeurIPS 2023Quantizable Transformers: Removing Outliers by Helping Attention Heads Do Nothing
• NeurIPS 2026Accelerating LLM Pre-Training through Flat-Direction Dynamics Enhancement
• ICML 2026LLM Priors for ERM over Programs
• ICLR 2026How Learning Rate Decay Wastes Your Best Data in Curriculum-Based LLM Pretraining
• CVPR 20263D-VCD Hallucination Mitigation in 3D-LLM Embodied Agents through Visual Contrastive
• NeurIPS 2025BrainEC-LLM: Brain Effective Connectivity Estimation by Multiscale Mixing LLM
• ICML 2025Position: Medical Large Language Model Benchmarks Should Prioritize Construct Validity
• ICLR 2025Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates

全部大模型基准论文

共 307 篇，第 6 / 7 页

251. NeurIPS 2026Meta-Learning Preferences for Multilingual LLM Alignment
252. NeurIPS 2026MinPath: Learning Efficient LLM Reasoning via Minimal Dependency Paths
253. NeurIPS 2026More Value per Key: Asymmetric Sparse Attention for Faster LLM Decoding
254. NeurIPS 2026Neglected Free Lunch from Post-training: Progress Advantage for LLM Agents
255. NeurIPS 2026Not All Gap Correction Helps: A Geometric View of External Information in LLM Inference
256. NeurIPS 2026Not All Proofs Are Equal: Evaluating LLM Proof Quality Beyond Correctness
257. NeurIPS 2026NuMuon: Nuclear-Norm-Constrained Muon for Compressible LLM Training
258. NeurIPS 2026NyoomFloat12: Accelerating LLM Inference via Lossless 12-bit Weight Compression
259. NeurIPS 2026On-Policy Consistency Training Improves LLM Safety with Minimal Capability Degradation
260. NeurIPS 2026OrchestraRL: Learning to Orchestrate LLM Agent Swarms with Entropy-Aware Communication Control
261. NeurIPS 2026Pacing Branch Parallelism in LLM Serving
262. NeurIPS 2026PCBInnoBench: Benchmarking LLM Agents on Real-World PCB Design
263. NeurIPS 2026Preserving Exploration for LLM Reasoning via Mean Order-Statistic Alignment
264. NeurIPS 2026PROBE: Learning to Audit Policy Compliance in Tool-Using LLM Agents
265. NeurIPS 2026Process-Aware LNS for Large-Scale MILP via Context-Enhanced Fine-Tuned LLM-driven Selector
266. NeurIPS 2026RA$^2$: Retain-Anchored Attraction Preserves Forget-Adjacent Utility in LLM Unlearning
267. NeurIPS 2026Rethinking Latency Denial-of-Service: Attack the LLM Serving Framework, Not the Model
268. NeurIPS 2026RIGOR: Risk-Gated Topology Adaptation for Robust LLM Multi-Agent Reasoning
269. NeurIPS 2026Salvation Lies Within: Proactive Prefix Re-forming for LLM-based Tagging
270. NeurIPS 2026SchemeArena: Factorized Stress Testing of Scheming in LLM Agents
271. NeurIPS 2026SCTI: Self-Calibrated Trident Identification of Black-Box LLM Watermarks
272. NeurIPS 2026Sequential Behavioral Watermarking for LLM Agents
273. NeurIPS 2026SFT-then-RL Outperforms Mixed-Policy Methods for LLM Reasoning
274. NeurIPS 2026SGEvolve: Semantic Gradient Guidance for LLM-Driven Evolutionary Search
275. NeurIPS 2026SkillMaster: Toward Autonomous Skill Mastery in LLM Agents
276. NeurIPS 2026Soteria: Formally Verified Planning with Runtime Enforcement for Safe LLM Agents
277. NeurIPS 2026Spatial-IQ: Deconstructing Spatial Intelligence via Hierarchical Capability Tests
278. NeurIPS 2026Specialists Hold, Generalists Discount: Asymmetric Equilibrium in LLM Routing Auctions
279. NeurIPS 2026Strategic Evaluation: Incentivizing AI Capability Coverage with Private Benchmarks
280. NeurIPS 2026STRIDE: Learnable Stepwise Language Feedback for LLM Reasoning
281. NeurIPS 2026Strong Teacher Not Needed? On Distillation in LLM Pretraining
282. NeurIPS 2026SudoBench: A Contextual Authorization Benchmark for LLM Agents
283. NeurIPS 2026SVDP: Training-Free Contextual Sparsity Predictors for Fast LLM Inference
284. NeurIPS 2026TextSeal: A Localized LLM Watermark for Provenance & Distillation Protection
285. NeurIPS 2026The Geometry of Forgetting: Temporal Knowledge Drift as an Independent Axis in LLM Representations
286. NeurIPS 2026The Illusion of Diversity: Aligning LLM Exploration via Effective Entropy
287. NeurIPS 2026The Labeling Problem in Hallucination Detection Benchmarks: An Empirical Evaluation
288. NeurIPS 2026The Scaling Paradox of Tool-Calling LLM Agents Under Realistic MCP Faults
289. NeurIPS 2026The Text Uncanny Valley: Non-Monotonic Performance Degradation in LLM Information Retrieval
290. NeurIPS 2026Tokens-per-Parameter Coverage Is Critical for Robust LLM Scaling Law Extrapolation
291. NeurIPS 2026Trust, but Don’t Verify: Epistemic Blind Spots in LLM Source Evaluation
292. NeurIPS 2026TSFM Meets LLM: Context as Covariate
293. NeurIPS 2026UMAS: System-Level Uncertainty Quantification for Multi-Agent LLM Systems
294. NeurIPS 2026Valid Best-Model Identification for LLM Evaluation via Low-Rank Factorization
295. NeurIPS 2026Values as Style: Disentangling Values from Semantics with One-Way Mixing for Low-Damage LLM Steering
296. NeurIPS 2026VEHBench: A Stage-Local Diagnostic Benchmark for LLM-Assisted Vibration Energy Harvester Design
297. NeurIPS 2026Verifiable LLM-Guided Focal SMT Solving for Quantified Arrays
298. NeurIPS 2026Verified-Source Authority Is Not Generic Sycophancy: Cue-Family Decomposition of LLM Compliance
299. NeurIPS 2026VeriVul: A Verification-Guided Framework for Generating Realistic Vulnerability Benchmarks
300. NeurIPS 2026ViroGym: Realistic Large-Scale Benchmarks for Evaluating Viral Proteins

常见问题

2023–2026 年四大会议一共有多少篇大模型基准论文？

共 544 篇，按年份分别是 2023 年 2 篇、2024 年 61 篇、2025 年 174 篇、2026 年 307 篇。

NeurIPS 2026 有多少篇大模型基准论文？

NeurIPS 2026 有 138 篇大模型基准论文，上一届是 92 篇。

ICLR 2026 有多少篇大模型基准论文？

ICLR 2026 有 103 篇大模型基准论文，上一届是 50 篇。

ICML 2026 有多少篇大模型基准论文？

ICML 2026 有 60 篇大模型基准论文，上一届是 24 篇。

CVPR 2026 有多少篇大模型基准论文？

CVPR 2026 有 6 篇大模型基准论文，上一届是 8 篇。

哪个会议的大模型基准论文最多？

NeurIPS，2023–2026 年共 256 篇。