Hybrid Reasoning Agents: LLM agents that think in latent space, act through text(arxiv.org)

497 points by allinonai 7 days ago | 231 comments

arXiv paper introducing HybridAgent, LLM agents that switch between latent-space reasoning and text-based tool interaction, trained with HybridSFT and HybridRL.

Paper2Agent: Stanford Reimagining Research Papers as Interactive AI Agents(arxiv.org)

134 points by Gaishan 1 year ago | 34 comments

Absolute Zero: Reinforced Self-Play Reasoning with Zero Data(arxiv.org)

63 points by leodriesch 1 year ago | 15 comments

Ring-Zero: Scaling Zero RL to a Trillion Parameters for Emergent Reasoning(arxiv.org)

51 points by binyu 3 months ago | 20 comments

Coding agents think ahead of time(arxiv.org)

79 points by andre15silva 3 months ago | 60 comments

Sycophantic AI Decreases Prosocial Intentions and Promotes Dependence (2025)(arxiv.org)

140 points by robin_reala 2 months ago | 77 comments

LLMs are making formal software verification feasible at a much larger scale(anthropic.com)

429 points by johnbender 3 days ago | 265 comments

Anthropic research post describing how Claude produced formal proofs of crash-freedom for RISC-V binaries of 123 OSS-Fuzz open-source libraries, finding bugs along the way.

Some critical issues with the SWE-bench dataset(arxiv.org)

331 points by joshwa 1 year ago | 113 comments

arXiv paper introducing SWE-Bench+, an empirical analysis revealing data quality issues in the SWE-bench coding benchmark for LLMs, such as solution leakage and weak tests.

Borrow Checking Zero-Cost Cyclic Mutability(antelang.org)

35 points by jfecher 7 days ago | 2 comments

A blog post from the Ante programming language project describing a borrow-checking approach that allows zero-cost cyclic shared mutability.

Stealing Reasoning Traces from Proprietary LLM APIs(stolen-thoughts.com)

665 points by quantumgarbage 2 months ago | 295 comments

Project page for a research paper on extracting hidden reasoning traces from proprietary LLM APIs, by authors from MATS, ELLIS, Max Planck and Tübingen.

AI Agent Guidelines for CS336 at Stanford(github.com)

480 points by prakashqwerty 4 months ago | 148 comments

A CLAUDE.md file in Stanford CS336's assignment repository giving guidelines for AI coding assistants working with students on the course.

C++ patterns for low-latency applications including high-frequency trading(arxiv.org)

343 points by chris_overseas 2 years ago | 200 comments

arXiv paper presenting C++ design patterns and benchmarks for low-latency code in high-frequency trading, including a Disruptor pattern implementation.