推論系統

推論 runtime、量化、 serving、KV cache、成本與效能。

u-OPSD Trains Solely Through Internal Model Voting, Boosting Qwen3-8B’s Five-Benchmark Math Average by 10.7 Points

u-OPSD uses eight reasoning trajectories from the same model to produce a majority-vote pseudo-answer, then distills token distributions conditioned on that consensus into trajectories that disagree. Qwen3-8B’s average across five benchmarks in non-thinking mode rose from 43.57 to 54.31, though incorrect consensus and the cost of repeated sampling still limit its applicability.

2026-08-08

最新模型

BaKron Reduces Two-Sided Hessian Quantization to Cubic Complexity, With Its 8192-Square-Matrix Kernel 60× Faster Than YAQA

Using anti-diagonal parallelization and recursive divide-and-conquer, BaKron reduces total work from quartic to cubic complexity while preserving curvature information on both the input and output sides. In 2.81-bit experiments on Llama 3 8B, it improves perplexity over GPTQ, but total quantization time remains longer, and no public implementation is currently available.

2026-08-08

代理工具

690 Agent Skills Put to the Test: Hybrid Retrieval Reaches 73.5% hit@5, While Adding an LLM Knowledge Graph Lowers It by 11.2 Points

A new study finds that pre-restricting candidate edges using embedding nearest neighbors prevents a knowledge graph from reaching skills the retriever could not already find. BM25–vector fusion still leaves about a quarter of queries unresolved, but outperforms graph neighbors under the same token budget.

2026-08-08

最新模型

Hyper-ES First Uses Gradients to Build a Low-Dimensional Search Space, Averaging 1% Higher on Mathematical Reasoning Than GRPO-LoRA

Rather than applying random evolution directly to billions of model-weight dimensions, Hyper-ES constructs a subspace from a small number of fine-tuning directions and then searches for layer-wise model-merging coefficients. The authors report modest gains across three models and six mathematical reasoning datasets, although the method is not entirely gradient-free.

2026-08-08

最新模型

MACRO Reorders Frozen Transformer Layers, Raising Average Accuracy by 5.0 Percentage Points Across Six Models

MACRO does not update model weights. Instead, it searches for a fixed execution path for each task that can skip layers, jump backward, and reuse hidden states. Across 13 benchmarks, the authors report reducing search time from 14.8 hours to 1.6 hours, though every new task still requires labeled training and validation data.

2026-08-08

最新模型

Google Releases WeatherNext 2 Code and Weights; Full-Model Inference Requires H100-Class Memory

Google DeepMind has released the JAX implementation, pretrained weights, and a streamlined Colab-compatible version of WeatherNext 2, covering global weather and tropical cyclone forecasts up to 15 days. Researchers can now reconstruct the autoregressive forecasting workflow, but the full model still depends on ECMWF initialization data and high-end accelerators, and its outputs cannot replace official warnings.

2026-08-08

推論系統

WitProbe Builds a Runtime Risk Ledger for Four Types of Attention Memory, With 12.4 Million Reads Staying Within Budget

A new study uses typed error contracts to unify monitoring across KV caches, latent caches, sparse selectors, and recurrent states, while using Lean to verify formalizable composition rules. The public repository can reproduce the paper’s figures and 67 theorems, but excludes architecture-specific probes and the complete serving platform.

2026-08-08

代理工具

CodeGrep Spins Repository Search Out Into a Dedicated 14B Agent, Using 19% Fewer Tokens on Successful Fixes

CodeGrep first narrows the scope of changes through parallel grep, glob, and file-reading operations, then passes candidate files to a frozen OpenHands coding agent. It maintains a roughly 27% repair rate on SWE-bench Verified while reducing inference costs, but the model and training pipeline have not yet been released.

2026-08-08

最新模型

RRC Converts Generative Reward Model Rankings into GRPO Signals, Raising AlpacaEval 2 Score from 35.8% to 41.3%

Instead of treating the probability of a preference token as the reward, RRC constructs scores from the relative rankings of candidate responses. With an 8B reward model and eight-vote aggregation, the ArenaHardV2 score also rises from 8.0% to 11.2%, at the cost of additional reward model inference.

2026-08-07

代理工具

Programmatic Tool Calling Matches or Beats JSON in 11 of 14 Models, With an 18.8-Percentage-Point Gap on Long-Chain Tasks

A new study compiles API schemas into typed Python stubs, allowing models to chain or invoke tools in parallel within a single inference. The method matches or outperforms native JSON for most models on a BFCL v4 subset, but it has so far been tested only with simulated tools that return their arguments, so the results do not establish that real-world API workflows would see the same benefits.

2026-08-07

最新模型

DASH Dynamically Assigns Distillation Weights Based on Reasoning Trajectories, Outperforming Fixed-Weight Baselines Across Three Qwen3 Model Sizes

Rather than treating teacher–student discrepancies at every token equally, DASH adjusts how far supervision can propagate based on the entire generation trajectory. The authors report consistent gains on three mathematics benchmarks and have released the training code, evaluation pipeline, and LoRA weights.

2026-08-07

代理工具

Argus Enables Fixed-Weight Agents to Accumulate Experience, Raising SWE-Bench Pro Accuracy From 59% to About 78%

Argus breaks long-running tasks into constrained missions, with a Manager, Planner, Engineer, and Reviewer jointly maintaining persistent state. Its authors report improving program repair performance at 1.41× the token cost; an agent-generated RWKV6 kernel was also merged into an open-source project after external corrections.

2026-08-06