Pan Lu

huggingface.co Paper page - Hybrid Reasoning Agents Join the discussion on this paper page

Jiaru "Rubin" Zou Verified account @Jiaru_Zou · Oct 14

Your LLM agent has to think out loud, token by token, before it can search, run code, or act in the world.

We humans don’t. We ponder in silence and speak when it matters.

Human language was never meant to be the limit of machine thought.

Introducing 🌗Hybrid Reasoning Agents
- Think in latent space, act through text
- Two modes, one unified multi-turn flow
- No black box, fully readable latent thoughts
- Any LLM can become a HybridAgent

View quoted post on X

Guowei Xu Verified account @Kevin_GuoweiXu · 22h

🚀 Can AI scientists come up with truly creative ideas?
A creative idea needs more than good performance. It must be original, useful and non-obvious.
Yet AI scientists are usually given as much task information as possible, which anchors them on familiar approaches.
We introduce BVSR: Blind Variation, Selective Retention, a discovery framework inspired by a classic theory of human creativity.
✅ Plugs into existing AI-scientist workflows.

View quoted post on X

elvis Verified account @omarsar0 · Sep 23

Must-read paper from Google on self-improving agent harnesses.

If you auto-optimize your agent's harness, your eval score can go up while the agent gets worse on real tasks.

This paper shows how to prevent that.

Of five harness-evolution methods compared on agentic workspace tasks, RRSI scored the lowest on the tasks it evolved against and highest on all three out-of-distribution benchmarks.

Automated harness evolution proposes edits to prompts, control flow, tools and memory, keeps the ones that raise the score, and repeats.

The authors show this overfits the training tasks. Meta-Harness reached 93.0 on the Harvey LAB evolve split but gained only 0.3 to 1.5 points on JobBench, GDPval and APEX-Agents.

RRSI adds regularization on both sides of the loop.

The proposer gets an edit budget that shrinks over time and is pushed toward directions it has not tried. A critic rejects benchmark-specific edits, and a pruner removes edits that are too small, too costly or no longer useful.

RRSI scored 90.5 on the evolve split and gained 3.5 to 4.7 points on the three held-out benchmarks. In the ablation, unregularized evolution used 3.80M tokens per trial against 2.42M for RRSI. With Gemini 3.5 Flash, RRSI raised Terminal-Bench 2.1 from 64.6 to 78.7 and carried a 2.2-point gain over to SWE-bench Verified.

Paper: arxiv.org/abs/2609.24972

Chat with Paper: academy.dair.ai/papers/rrsi-re…

View quoted post on X

Zhuofeng Li @ SGLang Summit Verified account @zhuofengli96475 · Sep 23

Thrilled to co-lead this effort with an incredible team!

What if agents take over thousands of GPUs and recursively improve themselves for days?

🚀 Introducing OpenRSI-Index (Preview v0.1): an open benchmark for Recursive Self-Improvement (RSI) in frontier model development.

Real research workflows. Original code & data. Fixed compute. Independent evaluation.

🧪 Signature Tasks: LLM pre-training, post-training & vision generation—50K H100-hours 🔥🔥🔥

🔁 RSI-Anything: idea → RSI task in ~1 hour. Automated review → build → run.

📣 CALL FOR CONTRIBUTORS: Bring an idea. RSI-Anything handles the rest in ~1 hour. Substantial contributions qualify for co-authorship.

Try it yourself!
🌐 Website: index.openrsi.foundation/index.html
🛠️ GitHub: github.com/OpenRSI-Founda…
🤝 Contribute: github.com/OpenRSI-Founda…

#RSI #OpenRSI #AutoResearch #Agents #OpenScience

View quoted post on X

OpenRSI Foundation Verified account @OpenRSI · Sep 23

As AI systems enter a recursive self-improvement loop, the central question is whether it can systematically move beyond human-designed methods to genuinely extend the scientific and intelligence frontier.

What’s missing is a neutral, open standard for evaluating these capabilities in real, production-scale intelligence development.

Today, we’re releasing OpenRSI-Index v0.1: evaluating whether AI can recursively improve itself and push the boundaries of intelligence and science, on production-scale clusters with 1k GPUs.

We turn fully open-source projects into autoresearch environments, with agent trajectories lasting 60+ hours. Building v0.1 took 100K+ H100-hours.

We’re building an ecosystem with and for the research community: let RSI benefit everyone, and let everyone shape RSI together.

We invite task contributors and compute partners to build this open benchmark with us - all contributors will be included as paper authors.
Shape RSI with us:
🌐 Website: index.openrsi.foundation/index.html
🛠️ GitHub: github.com/OpenRSI-Founda…
🤝 Contribute: github.com/OpenRSI-Founda…

View quoted post on X