Skip to content

Type: research platform

alphaXiv is a platform for discussing, discovering, and reading arXiv papers, surfacing trending work, threaded discussions, and author activity, and integrated into arXiv via arXivLabs. It adds a social and exploratory layer on top of the preprint corpus. GROUNDING tracks alphaXiv as a window into emerging AI research and community discussion.

Recent Updates

  • 2026-07-14: OS-Pruner: Pruning Chains-of-Thought of Reasoning Models via Optimal Stopping (cs.AI updates on arXiv.org) · arxiv.orgLLM Evals arXiv · arXiv.org · arXivLabs · CatalyzeX · Connected Papers · CORE · DagsHub · Gotit.pub · Hugging Face · Litmaps · scite · ScienceCast Aymane El Gadarri
  • 2026-07-14: LLM Jury Consensus Can Beat Self-Consistency for Reasoning Selection (cs.LG updates on arXiv.org) · arxiv.orgLLM Evals arXiv
  • 2026-07-14: Diagnosing Thinking Collapse in On-Policy Self-Distillation (cs.CL updates on arXiv.org) · arxiv.orgarXiv · Connected Papers · Litmaps · scite · DagsHub · Gotit.pub · Hugging Face · ScienceCast · CORE
  • 2026-07-14: Replicating Belief, Not Bits: Epistemic State Replication for Agentic Systems (cs.AI updates on arXiv.org) · arxiv.orgAgents arXiv · CatalyzeX · DagsHub · Gotit.pub · Hugging Face · ScienceCast · Connected Papers · Litmaps · scite · CORE
  • 2026-07-14: Constraint-Aware Hierarchical Search for regulation-driven fine-grained classification (cs.AI updates on arXiv.org) · arxiv.orgLLM Evals arXiv · arXivLabs · CatalyzeX · DagsHub · Gotit.pub · Hugging Face · ScienceCast · CORE · Connected Papers · Litmaps · scite
  • 2026-07-14: alphaXiv shares an agent workflow for reproducing paper experiments and building notebooks (alphaXiv) · twitter.comAgents
  • 2026-07-14: alphaXiv highlights a runnable Autoresearch workflow for a paper on memorized knowledge and finetuning (‌alphaXiv) · alphaxiv.orgAgents Tool Use
  • 2026-07-15: ICML 2026 Agent-Powered Paper Reproducibility Challenge (alphaXiv) · huggingface.coAgents Hugging Face
  • 2026-07-16: Diagnosing and mitigating context rot in long-horizon search (@askalphaxiv) · x.comAgents Context Engineering Long Context Fudan University Shanghai Jiao Tong University Shijie Xia Yikun Wang Zhen Huang Pengfei Liu GPT-OSS 120B · Qwen3.5-397B-A17B · GLM-4.7 GLM-5.0 · MiniMax-M2.5
  • 2026-07-20: RoboTTT claims long-context robot policy scaling with fast weights (alphaXiv) · x.comLong Context
  • 2026-07-20: On-Policy Delta Distillation focuses training on reasoning changes (alphaXiv) · x.comQwen3 · Gemma4
  • 2026-07-18: Hybrid SWA production results for long-context inference (alphaXiv) · x.comLong Context Xiaomi · Mimo v2.5
  • 2026-07-21: Blog post on world modeling for language agents (alphaXiv) · cameronrwolfe.substack.comAgents cwolferesearch cameronrwolfe
  • 2026-07-21: alphaXiv on recursive LMs as compositional generalizers (alphaXiv) · x.comAgents Context Engineering Tool Use
  • 2026-07-21: alphaXiv introduces OpenResearch for parallel research agents (alphaXiv) · x.comAgents Tool Use
  • 2026-07-21: OpenResearch CLI adds local research-agent workflows with parallel runs and compute backends (‌alphaXiv) · github.comAgents Tool Use OpenResearch GitHub · Claude Code · Codex · OpenCode · Modal · Hugging Face · Kubernetes · Slurm Weights & Biases
  • 2026-07-21: LOTUS paper claims faster CoT-like reasoning with looped Transformers (alphaXiv) · x.comLLM Evals LOTUS
  • 2026-07-21: Pretraining loss may predict RL gains in reasoning models (alphaXiv) · x.comLLM Evals
  • 2026-07-22: alphaXiv announces OpenResearch for running experiments on your own code and compute (alphaXiv) · x.comAgents Tool Use Code Agents askalphaxiv wandb
  • 2026-07-22: Reproduction report challenges RLSD’s headline accuracy claim (‌alphaXiv) · github.comLLM Evals Tinker API EasyVideoR1 veRL Ray · vLLM · Qwen-VL · marimo wandb · Qwen3-8B · Qwen3.5 4B
  • 2026-07-22: ActiveVision benchmark tests repeated visual observation and reasoning (alphaXiv) · x.comLLM Evals X JiaruiZ58876329 · Fable 5
  • 2026-07-24: ActiveVision benchmark shows multimodal models struggle with multi-step visual inspection (alphaXiv) · x.comLLM Evals OpenAI · Anthropic · GPT 5.5 · Claude Fable 5
  • 2026-07-24: Autonomous Topology Mutation for multi-agent LLM systems (cs.AI updates on arXiv.org) · arxiv.orgAgents Tool Use arXiv · Connected Papers · Litmaps · scite · DagsHub · Gotit.pub · Hugging Face · ScienceCast · CORE · arXivLabs · DeepSeek-V3
  • 2026-07-24: FlowEdit proposes information-theoretic control for branch-aware LLM reasoning (cs.AI updates on arXiv.org) · arxiv.orgLLM Evals arXiv · Connected Papers · Litmaps · scite · DagsHub · Gotit.pub · Hugging Face · ScienceCast · CORE
  • 2026-07-27: OpenResearch: Autonomous Research Agents with Parallel Experimentation (‌alphaXiv) · github.comAgents Code Agents Hugging Face · Modal

FAQ

What is alphaXiv?

alphaXiv is a platform for discussing, discovering, and reading arXiv papers, surfacing trending work, threaded discussions, and author activity, and integrated into arXiv via arXivLabs. It adds a social and exploratory layer on top of the preprint corpus. GROUNDING tracks alphaXiv as a window into emerging AI research and community discussion.

What does this page track?

Dated radar mentions, source links, related concepts, and builder-relevant context for alphaXiv, collected automatically by GROUNDING.

When was alphaXiv last mentioned?

alphaXiv was most recently mentioned in a radar update dated 2026-07-27.