NeoHorse-1: Towards Recursive Self-Improvement via Agentic Post-Training with Routing Harness

• URL: http://arxiv.org/abs/2609.08183v1
• Date: Tue, 08 Sep 2026 03:14:54 GMT
• Title: NeoHorse-1: Towards Recursive Self-Improvement via Agentic Post-Training with Routing Harness
• Abstract summary: NeoHorse-1 is a family of agent-native models developed to explore this path through agentic post-training.<n>Our system combines a heterogeneous model pool with intelligent routing, recording the predicted capability demand, selected service tier, and subsequent interaction for each user turn.
• Score: 76.6324710887112
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recursive self-improvement (RSI) requires a concrete mechanism through which an AI system observes its capabilities and converts that evidence into the next round of learning. We present NeoHorse-1, a family of agent-native models developed to explore this path through agentic post-training. Our system combines a heterogeneous model pool with intelligent routing, recording the predicted capability demand, selected service tier, and subsequent interaction for each user turn. These records are converted into training examples that preserve interleaved reasoning, tool calls, and harness context, and are admitted through structural validation, six-dimensional semantic evaluation, and subscene-level labeling. Routing signals organize supervised fine-tuning into a three-stage curriculum and extend to routing-guided on-policy distillation, where a teacher supervises student-generated responses under the same progression. Capability-guided allocation then converts evaluation feedback into the next training mixture, closing an evaluation-selection-update loop in which what the system learns to do shapes what it learns from next. Across eleven benchmarks covering harness-based agents, tool use, coding, and instruction following, post-training raises the macro-average from 58.94 to 64.87 at 4B and from 65.60 to 69.04 at 9B, substantially narrowing the aggregate gap between the post-trained 4B model and the 9B base model. NeoHorse-1 provides an initial prototype of this feedback-driven process and a path toward harness-mediated RSI across successive iterations.

Related papers

• A-SR: Self-Evolving Agentic LLMs for Symbolic Regression via Hierarchical Coordination [51.06539604709775]
  We propose A-SR, a self-evolving agentic framework that shifts the control unit from expression edits to role-conditioned evidence views.<n>A-SR coordinates formula discovery through routing among coordination protocols, an online evaluator-reward role policy, and state-routed process memory.
  arXiv  Detail & Related papers  (2026-08-05T14:01:10Z)
• Dynamic Mixture of Latent Memories for Self-Evolving Agents [57.20419419302731]
  MoLEM is a generative mixture of latent memory framework based on a dynamic mixture-of-experts (MoE)<n>We train the framework on continual-learning sequences spanning math, science, and code domains.<n>After the full continual-learning sequence, our method improves the average accuracy by 10.40% over the Vanilla pretrained baseline.
  arXiv  Detail & Related papers  (2026-05-21T03:35:10Z)
• Peer-Predictive Self-Training for Language Model Reasoning [15.499995918997756]
  We propose a label-free fine-tuning framework in which multiple language models improve collaboratively.<n>Given a prompt question, the models generate responses sequentially; the final aggregated answer serves as an internal target for learning.<n>On mathematical reasoning benchmarks, PST improves exact-match accuracy by 2.2 to 4.3 percentage points across Gemma-2-2B, LLaMA-3.2-1B, and Qwen-2.5-1.5B.
  arXiv  Detail & Related papers  (2026-04-14T23:29:44Z)
• Pioneer Agent: Continual Improvement of Small Language Models in Production [6.452260317191361]
  Small language models are attractive for production deployment due to their low cost, fast inference, and ease of specialization.<n>We present Pioneer Agent, a closed-loop system that automates the process of adapting small language models to a specific task.
  arXiv  Detail & Related papers  (2026-04-10T18:13:09Z)
• Mapping fNIRS Signals to Agent Performance: Toward Reinforcement Learning from Neural Feedback [2.933736356368701]
  Reinforcement Learning from Human Feedback (RLHF) is a methodology that aligns agent behavior with human preferences by integrating human feedback into the agent's training process.<n>We introduce a possible framework that employs passive Brain-Computer Interfaces (BCI) to guide agent training from implicit neural signals.<n>We present and release a novel dataset of functional near-infrared spectroscopy (fNIRS) recordings collected from 25 human participants across three domains: a Pick-and-Place Robot, Lunar Lander, and Flappy Bird.
  arXiv  Detail & Related papers  (2025-11-17T00:21:46Z)
• STARec: An Efficient Agent Framework for Recommender Systems via Autonomous Deliberate Reasoning [54.28691219536054]
  We introduce STARec, a slow-thinking augmented agent framework that endows recommender systems with autonomous deliberative reasoning capabilities.<n>We develop anchored reinforcement training - a two-stage paradigm combining structured knowledge distillation from advanced reasoning models with preference-aligned reward shaping.<n>Experiments on MovieLens 1M and Amazon CDs benchmarks demonstrate that STARec achieves substantial performance gains compared with state-of-the-art baselines.
  arXiv  Detail & Related papers  (2025-08-26T08:47:58Z)
• Optimizing Large Model Training through Overlapped Activation Recomputation [24.28543166026873]
  We present Lynx, a new recomputation framework to reduce overhead by overlapping recomputation with communication in training pipelines.<n>Our comprehensive evaluation using GPT models with 1.3B-23B parameters shows that Lynx outperforms existing recomputation approaches by up to 1.37x.
  arXiv  Detail & Related papers  (2024-06-13T02:31:36Z)
• Self-Supervised Representation Learning from Temporal Ordering of Automated Driving Sequences [49.91741677556553]
  We propose TempO, a temporal ordering pretext task for pre-training region-level feature representations for perception tasks. We embed each frame by an unordered set of proposal feature vectors, a representation that is natural for object detection or tracking systems. Extensive evaluations on the BDD100K, nuImages, and MOT17 datasets show that our TempO pre-training approach outperforms single-frame self-supervised learning methods.
  arXiv  Detail & Related papers  (2023-02-17T18:18:27Z)
• Automatic Recall Machines: Internal Replay, Continual Learning and the Brain [104.38824285741248]
  Replay in neural networks involves training on sequential data with memorized samples, which counteracts forgetting of previous behavior caused by non-stationarity. We present a method where these auxiliary samples are generated on the fly, given only the model that is being trained for the assessed objective. Instead the implicit memory of learned samples within the assessed model itself is exploited.
  arXiv  Detail & Related papers  (2020-06-22T15:07:06Z)
• Semi-supervised ASR by End-to-end Self-training [18.725686837244265]
  We propose a self-training method with an end-to-end system for semi-supervised ASR. We iteratively generate pseudo-labels on a mini-batch of unsupervised utterances with the current model, and use the pseudo-labels to augment the supervised data for immediate model update. Our method gives 14.4% relative WER improvement over a carefully-trained base system with data augmentation, reducing the performance gap between the base system and the oracle system by 50%.
  arXiv  Detail & Related papers  (2020-01-24T18:22:57Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.