Research and tools are moving toward agents that can be controlled, orchestrated and improved without retraining foundation models. Multi-agent topology selection, self-improvement loops, hybrid reasoning and calibration agents are emerging use cases.
A study on wake simulations and damped oscillators shows that physics-labeled LLM prompts improve initial calibration in 23 of 25 trials, but the advantage fades after 15 evaluations. Hybridizing with a classical optimizer cuts runs to 9, versus 17 f...
Researchers propose Structured Authority Injection, a protocol that routes human corrections through typed, schema-validated MCP tool calls and maintains a ranked authority context prepended to each LLM prompt. It defines three override classes with ...
Google has open-sourced Harness2, a system that lets an AI agent refine its own harness for each task and retain lessons learned, without touching model weights. Tested with Gemini 3.5 Flash and GLM-5 on legal, office, and professional benchmarks, it...
A preprint from Stanford, Together AI, UIUC, and Meta introduces Hybrid Reasoning Agents: models that reason internally and emit text only for tool use and final answers. In reported tests they gain about 10%, use half the tokens, run twice as fast, ...
PRISM recovers per-query signal from execution-verified graph corpora that lack rankings. A teacher reads a sibling graph via a zero-initialized low-rank branch; the student aligns per-decision targets on self-sampled trajectories. Across six benchma...
The FastAPI platform described by DeepUnfolding Agent separates adaptive control from domain-specific execution. Executors register a schema, send round-level diagnostics, receive schema-bounded actions, and return feedback. SQLite persistence and Do...
A 9-billion-parameter model, trained once on 2,506 conversations generated from 297 procedures, faces unseen debt collection procedures. Documented checks rise from 67-71% to 93-100%, while grounded checks already pass. An audit of the shared specifi...
Researchers use an inverse LLM to reconstruct reasoning traces from answers alone. Trained once, the model generates intermediate steps offline for prompt-answer data, avoiding the multiple rollouts of reinforcement learning. On Qwen3-0.6B, a single ...
