SURL AI News
繁中English日本語

Archive

All News

1537 technical news articles

推論系統

ThinkRetrieve Retrieves Worked Examples Mid-Reasoning, Boosting Qwen3-1.7B’s AIME 2025 Accuracy by 13.4 Points

Rather than merely asking models to think longer, ThinkRetrieve retrieves complete solutions to similar problems at each reasoning boundary and inserts them into the next segment of the reasoning context. All five tested models, ranging from 1.5B to 8B parameters, outperformed sequential test-time compute, but the method requires a large, rigorously decontaminated solution database.
最新模型

Adding Compatible Adjectives Still Causes SAEs to Lose 20% to 60% of Activated Features, Challenging the “Bag of Features” Intuition

A Heidelberg University study finds that although individual latents in sparse autoencoders may be interpretable, the full set of activated latents does not map reliably onto human concepts. Simply adding a compatible adjective before a noun can make 20% to 60% of the original activation set disappear, limiting the reliability of interpreting model states through set overlap.
最新模型

Muse Glimmer 30B Brings Multimodal Agents to a Single GPU With 17GB Quantized Weights and a DFlash Draft Model

Meta has released the Muse Glimmer 30B weights, vision encoder, and DFlash draft model. The 17GB version can run text, image, and speculative decoding workloads simultaneously within 24GB of VRAM. Meta measured an average of 233.4 tokens/s on an RTX 5090, but safety evaluations and deployment details show that the model still requires external safeguards and a newer runtime.
AI 輔助科學研究

Dual-Agent Research System Derives a New Lower Bound for the Grothendieck Constant From Repeated Failures, but the Key Pivot Still Came From Humans

Using reasoning models, coding agents, and file-based memory across roughly 240 research sessions, a team ultimately proposed—and the authors verified—`K_G ≥ 6π/11`. The system demonstrated strong local proof capabilities but did not independently recognize that its search had reached a bottleneck; research direction and decisions about which results to pursue still depended on roughly 40 human instructions.
代理工具

ChemWorld Compiles Chemical Processes Into Replayable Agent Environments, Preserving Transaction Receipts Even for Failed Actions

ChemWorld uses typed operations, a compatibility compiler, and private chemical laws to let researchers replace a single world rule while retaining the same public interface. The open-source release passed 52 generated combinations and eight end-to-end workflows, but it validates software execution semantics—not real-world chemical accuracy.
最新模型

UniF-MoE Extracts Shared Blocks Before Routing Residual Computation, Cutting Inference Time by 45.2% Versus Top-2 GMoE

Rather than deciding shared experts, FFN width, and the number of activated experts separately, UniF-MoE lets each token first select shared channels and then accumulates routing probabilities according to its remaining requirements. In the authors’ DeiT and BERT experiments, it improves accuracy and MoE inference cost, but it has not yet been validated on generative large language models or distributed expert parallelism.
代理工具

Cross-Language Prompts Alter Agent Tool Paths, With Four Large Models Retaining Only 71%–73% of Their Action Policies

A Microsoft research team compared 41 languages across 2.38 million rollouts and found that agents take different tool paths when the same task is presented in another language—even with greedy decoding. The study also shows that non-English tasks rely heavily on first translating into English, although its tools are merely symbolic calls and do not capture the errors or side effects of real-world APIs.
最新模型

ReRound Uses a Diffusion Prior to Rechoose Quantization Rounding, Boosting Average Accuracy of 3-/4-Bit Small Models by Up to 1.6 Points

ReRound uses no calibration corpus. Instead, it trains a conditional diffusion model on the model’s own weights to reconsider whether weights near the midpoint of a quantization interval should be rounded up or down. It preserves the existing low-bit format at deployment with no additional inference overhead, but each model requires several hours of model-specific reconstruction upfront.