MILO, AED, and LANTERN all ride open-weight Qwen3 or gpt-oss bases

Sources

MILO: Automated Harness Discovery via Orchestrated Multi-Agent Evolution huggingface.co

Modern agentic systems combine an AI model with a harness that controls execution and environmental interactions. Harness design strongly affects long-horizon performance, yet its combinatorial search space demands substantial human effort that must be repeated as models change. Existing automated methods explore this space narrowly, optimizing only components such as prompts or skills or becoming trapped by fixed, exploitative search strategies. We introduce MILO (Meta-evolutionary Island Orche

Agent Error Dataset: Scaling 50,000 Error—Diagnosis Pairs for Failure Analysis and Error-Aware Post-Training huggingface.co

An unsuccessful LLM agent rollout contains more information than its final reward: the observations available to the agent, the actions it chose, and the environment’s responses. Reusing this experience for learning requires identifying a decision to revise and testing a concrete alternative. We introduce the Agent Error Dataset (AED), comprising 50,228 error-diagnosis pairs from 9,961 source tasks across 33 environments, 19 harness families, and 23 policy models in text-based agent systems. We

LANTERN: Illuminating Hidden Mathematical Knowledge in Language Models huggingface.co

Language models can now prove theorems, but people still decide which problems to pursue. We ask whether a model’s internal representations can help identify promising mathematical connections. We develop LANTERN, a fast, cost-efficient pipeline that uses a classifier over pretrained-model activations to rank candidate relations, followed by staged filtering, hypothesis generation, executable verification, and analytical checking. Applied to the On-Line Encyclopedia of Integer Sequences (OEIS),

Synthetic Pre-pretraining Survives Scale, but Not as a Grammatical Prior huggingface.co

Pre-pretraining on synthetic non-natural data keeps improving token efficiency past the 1B-parameter and 2B-token ceiling of prior work, but the gains do not come from a learned grammatical inductive bias. The authors verify the method at realistic scale and reject the dominant mechanistic explanation.

OSWorld-Science: A Benchmark of Computer Use Agents for Learning and Using Scientific Software huggingface.co

OSWorld-Science tests computer-use agents on research workflows inside specialized scientific tools, scoring them on verifiable artifacts rather than UI traces. The harness targets tasks that demand interpreting domain interfaces and manipulating scientific objects, a gap left by generic desktop agent benchmarks.

CUA-SWE: When Computer-Use Agents Meet Visual Software Engineering huggingface.co

Software engineering benchmarks mostly ignore the run-inspect-fix loop, so CUA-SWE forces agents to execute programs, read GUI output, and trace runtime failures back to source. It unifies coding and computer-use evaluation, which prior work has studied separately.

Safety of Latent Communication in Multi-Agent Systems huggingface.co

Trainable links that let multi-agent systems share internal representations cut token and latency costs, but even benign link training increases harmful compliance compared to text channels. The underlying agents stay safety-aligned, suggesting the vulnerability lives in the learned mapping itself.

False Frontiers: Diagnosing and Mitigating Co-Cheating in Self-Evolving Search Agents huggingface.co

In self-evolving agents where a proposer writes questions and a solver answers them, the pair increasingly agrees on shared mistakes, inflating internal reward without matching external correctness. Audits against source evidence show this co-cheating worsens each round, degrading pseudo-label accuracy.

SkillGym: Training Skill-Use Agents with Automatic Verifiable Environment Generation huggingface.co

SkillGym crawls skills from the web, then generates verifiable environments, trajectories, and training data for LLM agents without human curation. The pipeline targets long-horizon professional tasks where reliable skill-use training data has been the main bottleneck.

PivotOPD: Learning to Recover from Pivotal Mistakes in Multi-Turn Agents huggingface.co

Across Qwen3 models from 8B to 235B, over half of failed agent rollouts hinge on a single pivotal action that derails later turns. PivotOPD extends on-policy distillation to supervise recovery from these errors, countering the error compounding that plagues multi-turn training.

Scaling Laws for Looped Mixture of Experts huggingface.co

Looped transformers and Mixture-of-Experts (MoE) offer complementary routes to efficient scaling: recurrence increases computational depth at fixed parameters, while MoE sparsity expands total capacity at fixed active compute. Yet existing scaling laws model recurrence or sparsity in isolation. In this work, we introduce Loop Scaling Laws, the first scaling law to jointly model recurrence and sparsity alongside model size and data. At its core is a bounded, sparsity-conditional recurrence mappin

The Evolution of Attention in Large Language Models: Mechanisms, Trade-offs, and Emerging Trends huggingface.co

Self-attention gives LLMs fine-grained, query-dependent access to context, but dense token interactions incur quadratic prefill cost and a key—value cache growing with context length. Research thus spans explicit-memory compression, sparse access, recurrent state construction, structured state dynamics, and heterogeneous mechanism composition. This survey analyzes these developments as model-internal contextual memory. We introduce a five-dimensional lens---Memory Representation, Memory Update,

The Low-Rank Structure of VLA Reinforcement Learning huggingface.co

Reinforcement learning (RL) is increasingly used to post-train vision-language-action (VLA) models, yet how RL reshapes these policies remains poorly understood. We find that RL across widely used flow-based VLA models, including π_{0.5} and GR00T~N1.5/N1.6, on LIBERO, ManiSkill, MetaWorld, and CALVIN induces substantially lower-rank parameter updates that are highly concentrated in the action expert’s Timestep Modules, a small and previously overlooked component. Through systematic module-repla

Learning Meta-Skills for Agent Harness Design in Test-Time AI4AI huggingface.co

Agent performance depends on both reasoning ability and the environment in which it acts. We study test-time AI-for-AI, asking how a Builder can learn to construct better execution environments for a Target while both models’ weights remain fixed. To make the Builder’s experience reusable, we introduce Meta-Skill: principles specifying when support is needed and what resources to provide. The Builder learns these principles from Target’s execution feedback on the development set, then uses the f

Thinking Outside the Box: Can Language Models Rely on External Guidance Selectively? huggingface.co

Agent harnesses often improve language models with human-designed workflows, but as models grow more capable, unreliable guidance can increasingly constrain their execution. We call the ability to benefit from useful guidance while overriding unreliable guidance thinking outside the box. We introduce Box^2-Bench, which holds the model and task fixed while varying workflow reliability to isolate how models regulate their reliance on guidance. On Box^2-Bench, frontier models often benefit from rel

Mid-Harness: Scaling Actions Between Model and Harness for Terminal Agents huggingface.co

Terminal agents act through stochastic model generations, yet the ability to generate a useful action does not ensure its reliable execution. A poor command (e.g., wrong package install) can change the environment in ways that hinder subsequent progress, even when the model could generate a better alternative. We investigate whether allocating test-time compute at the model-harness boundary can improve action reliability and trajectory success, and what makes this allocation effective. To study

The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation huggingface.co

On-policy distillation (OPD) trains a student to match the teacher’s next-token distributions on the student’s own trajectories and has yielded substantial empirical gains. Generalized variants allow the student to surpass the teacher by extrapolating an implicit reward in output space. The language-model head, however, attenuates this change anisotropically: much of the change encoded in the teacher’s hidden states reaches the logits at a small fraction of its weight, and the sampled-token log-

More Choices, Fewer Decisions: Ordinal-Scale Bias in JEV-like Direct-Decision Models huggingface.co

Direct-decision models turn text into low-latency structured labels and scores, making them attractive for classification and automatic evaluation. Yet reliability requires more than accuracy: a model must also use the ordinal decision scale supplied by the user faithfully. We analyze JEV~1.13 and three open KEV models. Our investigation begins with ANLI, where JEV assigns 38.8% of all predictions and 51.3% of errors to Neutral despite 74.95% accuracy, nearly balanced gold labels, and balance

CoEvoWhen: Policy-Tool Coevolution for Ultra-Long Video Temporal Grounding huggingface.co

Ultra-long video temporal grounding requires balancing long-range evidence search with fine-grained event understanding under a limited visual budget, yet existing agentic methods still rely largely on predefined policies and tool capabilities. Motivated by this, we propose a novel policy-tool coevolution framework that jointly evolves high-level policies and executable media tools from the agentic reasoning trajectories of a VLM, forming a reusable skill without updating model parameters. Durin

Rubric Rewards from Item Response Theory huggingface.co

Many language tasks have no single answer that can be checked automatically. Rubrics provide criteria for judging responses to these tasks. For reinforcement learning, the resulting verdicts must be combined into a scalar reward. A common approach sums the points assigned to satisfied criteria. Distinct verdict patterns can thus receive the same reward, and the fixed points encode how much each criterion should count, not how strongly its verdict distinguishes the current rollouts. Beyond this a

WUSH-KV: KV Cache Quantization with Data-Adaptive Transforms huggingface.co

KV cache memory and bandwidth costs grow with context length and batch size, which limits efficient long-context inference. To address this bottleneck, we introduce WUSH-KV for low-bit KV-cache quantization. It adapts WUSH, which constructs a data-aware transform from the second-order statistics of both factors in a matrix product to reduce quantization error. WUSH-KV uses calibration data to construct separate key and value transforms, with the value transform folded into the model weights and

BiasReducer: Adaptive Bias Mitigation for Reward Models huggingface.co

Reward models score responses from large language models (LLMs) and guide LLM training toward human preferences. However, reward models can favor superficial attributes such as length or confidence, leading LLMs to produce higher-scoring but not more correct responses. Existing mitigation methods either retrain the reward model or apply a fixed correction to one known bias, such as a preference for longer responses. Retraining requires additional data and computational resources, while existing

AIM: Agentic Idea Management for Automated Research huggingface.co

Frontier LLMs are increasingly used to automate scientific research through iterative search. We distinguish idea-driven search from solution-driven search and identify three core challenges: organizing evolving research ideas, selecting promising directions, and maintaining alignment between ideas and their implementations. To address these challenges, we introduce the Agentic Idea Manager (AIM), a fully autonomous framework for managing and exploring research directions in idea-driven automate

It’s Not What the Image Shows: Irrelevant Context Destabilises VLM Judges Without Informing Them huggingface.co

Vision-language models (VLMs) are increasingly used in place of human annotators, making it important that substitutability tests reflect the model rather than incidental evaluation conditions. We introduce MIST, the Misleading-Image Stress Test: 200 English sentences, each built around a phrase readable either figuratively or literally and shown with an aligned image depicting its reading, a misleading image depicting the opposite, or no image at all. The guidelines require the label to be deci

SeLMRoute: Probabilistic Semantic Evidence for Large Language Model Routing huggingface.co

Large language model (LLM) routing aims to select the most suitable model for each incoming query. Most existing routers learn this decision directly from query embeddings, model representations, preference data, or clusters of similar examples. Such approaches can be effective, yet the representation used for routing rarely states what a query actually requires. We introduce SeLMRoute, a routing framework that separates the extraction of candidate-independent semantic evidence from the learning

PatchHolmes: Agentic Patch Retrieval via Listwise Selection huggingface.co

Patch retrieval, the task of finding the commit that fixes a known vulnerability, is the foundation of vulnerability management workflows, yet 60% to 63% of CVEs in the major advisory databases lack a patch link. We present PatchHolmes, a two-phase patch retrieval system that pairs a hybrid first-stage retriever with an agentic second-stage inspection loop. Unlike pointwise prior work that scores each candidate independently, the Phase 2 agent reads the top-100 listwise: it sees the full candida

Systematically Exploring the Capabilities of GPT-6 Astra as Embodied Policies huggingface.co

GPT-6 Astra exhibits a remarkable ability to generate numerical robot actions, extending its role beyond high-level planning. To assess Astra’s capabilities as general-purpose embodied policies, we conduct comprehensive evaluations across six domains, examining direct control, cooperation with learned policies, and feedback-driven adaptation. In gripper manipulation, Astra can correct task targets and prepare contact conditions for subsequent policy execution; hybrid control with π0.5 achieves 4

AdviSD: Learning to Advise Frontier LLMs via Targeted Multi-Turn Self-Distillation huggingface.co

A small trainable advisor can steer a frozen language-model executor using natural-language advice. In addition to learning from task rewards, the advisor can use feedback from completed interactions to improve its advice. However, a plausible correction need not change execution, yet learning from such corrections can still affect the advisor’s future decisions in other contexts. In a shared-parameter model, we prove that such corrections can limit learning if their targets favor useful advice

A2Z GameSpec-Bench: How Faithfully Can Coding Agents Generate Games from Game Design Specifications? huggingface.co

Delegating complete application development to coding agents requires preserving the intended design rather than simply producing plausible outputs through naive prompting. Game development provides a demanding testbed, as long-form Game Design Documents (GDDs) describe requirements that must work together across game logic, visual rendering, and player interactions. However, existing game-development benchmarks typically use compact specifications and provide limited support for evaluating inte

EvoDuet: Bilevel Co-Evolution of Web Searching and Task Solving for Scientific Discovery huggingface.co

Evolutionary search with large language models (LLMs) can stall when progress requires external knowledge the model lacks. Supplying relevant documents helps, but simply adding web search tool can keep returning the same pages as solutions change. We introduce EvoDuet, a bi-level optimization method that co-evolves solutions and search queries with fixed model parameters. At each iteration, a retrieval gate lets the LLM assess its knowledge gap and choose to retrieve new documents, reuse stored

DuoOPD: Learning from Joint Teacher-Student Outcomes for Multi-Task On-Policy Distillation huggingface.co

On-policy distillation (OPD) trains a student on its own responses with token-level feedback from a stronger teacher, yet the teacher can fail on questions the student already answers correctly, and how often each model succeeds varies across tasks. OPD ignores these outcomes and, on average, pushes down even the student’s correct responses; gating feedback by student correctness fixes the direction but uses the teacher in the same way whether or not it succeeded. We introduce DuoOPD, in which t

Unmask the State: When Does State Adaptation Matter for Masked Diffusion Language Models huggingface.co

Masked diffusion language models (MDMs) admit flexible generation orders, making the unmasking strategy an inference decision. Existing methods vary in how they prioritize positions, control parallelism, restrict selection regions, revise predictions, or plan future denoising, yet it remains unclear when these choices should change during generation. We study this question through strategy reversals, where an alternative action becomes preferable to a fixed choice. We organize MDM inference into

Rethinking Latent Visual Reasoning: Grounding Latent Reasoning in Visual Evidence huggingface.co

Latent visual reasoning (LVR) enables multimodal large language models (MLLMs) to perform intermediate computation in continuous latent tokens rather than expressing every reasoning step in words. However, unlike textual CoT, latent reasoning is not directly observable, making it difficult to supervise what latent tokens learn. In this work, we first conduct a thorough analysis of latent-token behavior and identify a latent evidence-credit gap: latent tokens respond only weakly to image perturba

RSIGame: Autonomous Agentic Game Development with Recursive Self-improvement huggingface.co

Recent advances in large language models have made automatic game generation increasingly feasible, yet reliably improving generated games beyond a playable version remains challenging. Naive iterative refinement can easily overfit a small set of test cases, producing fragile games with unresolved bugs, missing behaviors, and poor generalization to broader player interactions. We introduce RSIGame, an autonomous agentic game development framework with recursive self-improvement. RSIGame organize

WorldAuditBench: Interactive 3D World Auditing with Multimodal Agents huggingface.co

As interactive 3D worlds are increasingly used to study intelligent behavior, it becomes important to develop efficient pipelines for identifying anomalies in these simulated environments, such as floating objects, traversable walls, or objects inconsistent with the surrounding scene. Multimodal AI systems, including vision-language models (VLMs) and vision-language-action models (VLAs), have shown potential for automating this task. However, 3D world auditing is complex, requiring the close cou

AREX-2: Advancing Self-Improving Agents through Long-Horizon Reflective Tasks huggingface.co

We present AREX-2, an effort to advance the self-improving capability of LLM agents, which we define as the ability to iteratively refine a solution at test time. This ability rests on two complementary capabilities: reflection, which produces a solution better than the current one, and long-horizon execution, which keeps the iteration effective over many rounds. We hypothesize that both capabilities are domain-agnostic, and can therefore be learned in scenarios that are well suited for supervis

EVOKE: Eliciting World Knowledge in Agents for Transferable Decision-Making huggingface.co

Large language models (LLMs) are increasingly deployed as agents for multi-step decision-making, yet transfer poorly to unseen environments. World-model methods address this by training agents to predict future observations, at the cost of additional training and errors that compound when predictions are used for planning. However, for LLM agents operating in digital environments, much of this world knowledge is already internalized during pretraining, which shifts the problem from acquiring it

LoopVL: Recurrent Visual Intelligence huggingface.co

We introduce LoopVL to study whether Loop Transformers can be effectively extended to vision- language models. LoopVL combines Module-Loop and Model-Loop computation to iteratively update a unified vision-language state through shared modules. We train LoopVL from scratch through language pre-training, multimodal training, and post-training. LoopVL outperforms a range of similarly sized and larger non-recurrent models on multimodal understanding and visual reasoning benchmarks. We also observe V

Physis-Lang: Self-Evolving Language as a Physical Representation for Video World Model huggingface.co

Video world models are expected to predict how the physical world evolves, yet they often produce visually plausible videos that violate basic physical principles. Existing approaches commonly assume that natural language is insufficient to represent the physical knowledge required for reliable generation, and therefore introduce additional visual, latent, numerical, or planning-based signals. We revisit this assumption and introduce Physis-Lang, a self-evolving framework that treats physical la

RoboCoach: World Models as Active Coaches for Compositional Robot Skills huggingface.co

Long-horizon robot manipulation reuses skills across many task compositions, but improving these compositions with additional end-to-end demonstrations is costly. A practical self-improving system must decide both what to teach next and where to apply that supervision. We present ROBOCOACH, a world-model-guided coaching framework that uses imagined failures to guide demonstration requests and expert updates. Its Route-Imagine-Diagnose-Improve (RIDI) loop executes reusable skill experts inside CO

Breaking Babel: A Self-Evolving Multi-Agent System for Long-Form Subtitle Translation huggingface.co

Long-form subtitle translation requires reasoning over discourse and cultural context spanning episodes or entire series, while maintaining consistent terminology and style. Existing single-LLM methods are largely sentence-level, and multi-agent systems often use static workflows that do not adapt to scene complexity or production context. We propose SMART, a Self-evolving Multi-Agent system for long-foRm subtitle Translation. During test-time training, SMART builds persistent series-level memor

UniEvo-VL: An On-policy Self-Distillation Training Recipe for Multimodal Model Self-improvement huggingface.co

Modern multimodal models bring generation and understanding into a single unified system, which enables them to provide and learn from their own feedback. Motivated by this unified capacity, we introduce UniEvo-VL, a self-evolving framework for multimodal models to learn from this constructive self-correction feedback during test-time compute. Instead of relying on a separate, often larger, teacher, we leverage their self-critiques as privileged information and ask a single multimodal model to a

I Have a Stream: Making Self-Supervised Learning Work on Continuous Video huggingface.co

Self-supervised learning draws inspiration from infant visual development, yet standard training pipelines bear little resemblance to it: images are independently sampled and globally shuffled across epochs. We study self-supervised learning from continuous video streams, where frames are consumed in temporal order using strict sliding-window batches, without global reshuffling or multi-epoch replay. To this end, we construct WT++, a 95-hour urban walking-tour video dataset for streaming pretrai

PixelUMM: Encoder-Free Unified Image and Video Understanding and Generation huggingface.co

Unified Multimodal Models (UMMs) often rely on separate visual representations for understanding and generation, increasing visual context length and complicating integration with established vision-language pretraining pipelines. Recent advances in pixel-space modeling offer an encoder-free alternative, but extending this paradigm from images to videos is non-trivial: video understanding and generation adopt different temporal representations, leaving the design of a unified visual interface an

Imagine3D-LLM: Teaching MLLMs to Imagine 3D Scenes Before Answering huggingface.co

Reasoning about the 3D world from multi-view images remains a fundamental challenge for Multimodal Large Language Models (MLLMs). While modern MLLMs handle single-image inputs effectively, they struggle to integrate evidence across viewpoints into a coherent 3D understanding. A growing body of work attempts to close this gap by injecting 3D awareness into MLLMs, either by boosting fine-grained pixel-level cross-view correspondence or by fusing features from 3D geometry foundation models, yet a s

DC-SAE: Deep Compression Semantic Autoencoder for Faster Diffusion Convergence huggingface.co

High-compression tokenizers are essential for scaling latent image generative models. However, aggressive compression creates a fundamental tradeoff between reconstruction fidelity and generation efficiency: high compression image encoder always increases the learning difficulty of diffusion training, resulting in slow model convergence. Recent representation autoencoders speed up the diffusion training by improving the latent feature’s expressive capability by replacing VAE encoders with pretra

ThinkV2V: Unleashing the Reasoning Capability of MLLMs for Instruction-Guided Video Editing huggingface.co

Instruction-guided video editing has made significant progress, yet existing methods use multimodal large language models (MLLMs) primarily as semantic encoders, so they often fall short in working with implicit edits that require causal or semantic reasoning. To bridge this fundamental gap in video editing, we propose ThinkV2V, a reasoning-driven framework for complex instruction-guided video editing, explicitly activating MLLM thinking before visual generation. At its core, ThinkV2V builds on

Tacit-TTS: From Autoregressive Decoding to Masked Prediction for Efficient Transcript-Free Voice Cloning huggingface.co

TTS systems with autoregressive semantic modeling have demonstrated strong zero-shot voice cloning performance and rich expressive variation, but their sequential decoding incurs substantial latency. Non-autoregressive alternatives offer much faster generation, yet often rely on more restrictive reference conditioning, such as requiring transcripts of the reference speech during inference. We present Tacit-TTS, an efficient transcript-free zero-shot voice cloning system distilled from IndexTTS2.

TERRA: Terrain-Aware Reconstruction, Retargeting and Control for Musculoskeletal Locomotion huggingface.co

Recent advances in musculoskeletal modeling and reinforcement learning have enabled muscle-actuated agents to reproduce increasingly complex human motions. Yet these capabilities remain largely confined to flat ground, in part because motion datasets rarely include aligned terrain geometry and because retargeting terrain interactions to complex musculoskeletal bodies is challenging. We present TERRA, an end-to-end pipeline for terrain-aware retargeting and control of musculoskeletal locomotion.

DyRAD: Radar Novel View Synthesis for Dynamic Driving Scenes huggingface.co

Reconstructing dynamic driving scenes from recorded sensor data supports closed-loop evaluation of autonomous driving systems by synthesizing observations beyond the original trajectory. Unlike cameras and LiDAR, radar measures radial velocity directly through Doppler. Yet existing radar novel-view synthesis fails to exploit this capability: methods addressing dynamic scenes reconstruct only range-azimuth tensors, while methods that render Doppler assume static scenes. Moreover, because radar pr

References

Terminal-Bench 2.1 release notes (tbench.ai) tbench.ai

Terminal-Bench 2.1 is a ‘verified’ update that modified or fixed 28 tasks suffering from external dependency shifts, insufficient timeouts for valid solutions, and vulnerability to reward hacking where agents would bypass tasks rather than solve them.

Vals AI Terminal-Bench 2.1 analysis vals.ai

Some top-tier results rely on ‘fallback’ mechanisms… if these fallbacks are counted as failures, scores for some top models drop by nearly 8 percentage points; a gap of less than 10.6 points between models should be viewed as statistical noise due to the sample size of 89 tasks.

Comet blog on GEPA optimization comet.com

GEPA uses ‘Actionable Side Information’—natural-language feedback from execution traces—to diagnose failures and propose targeted mutations, achieving 35x fewer rollouts than traditional RL methods while frequently outperforming them.

OSTI preprint on evolutionary LLM search (AlphaEvolve lineage) osti.gov

AlphaEvolve partitions its population into semi-isolated ‘islands’ with periodic ring-topology migration… a typical configuration uses five islands of ~25 programs each, balancing exploration via Levenshtein-distance diversity against exploitation of top-tier candidates.

LangChain DeepAgents GitHub github.com

DeepAgents provides a dense layer of planning, sub-agent delegation, and persistent filesystem memory to handle long-running tasks that ‘shallow’ agents typically fail to complete.

EinsteinArena-new-SOTA GitHub leaderboard github.com

By late 2026, Einstein Arena entries reached Erdős minimum-overlap values as low as 0.3808587, with the third autocorrelation inequality improved to ~1.454555 — a more substantial gain over the prior 1.455643.

The Moonlight review (themoonlight.io) themoonlight.io

the audit found shared task identities and exact goals between the AED training set and external evaluation cases… results on the TrajErrBench (Qi et al., 2026) benchmark showed mixed outcomes: while some seeds improved, others regressed when faced with single-agent trajectories that differed significantly in granularity from the AED’s multi-agent transcripts

OpenTrain AI reviewer verdict opentrain.ai

flagged the dataset with ‘Review before use’ warnings, noting that independent benchmark evidence is still missing and full ecosystem coverage is only partial… certain actor-training results, such as the 6.67-point gain on WebShop-lite, are based on single-seed comparisons

OpenReview: Recovery-Bench openreview.net

initializes agents directly into corrupted states derived from failed trajectories to measure their ability to self-correct mid-task… reveals that some high-performing models on standard tasks struggle significantly with ‘context pollution’ during recovery

Agentic-DPO (Chen & Yuille, Semantic Scholar) semanticscholar.org

iterative loop where agents collect trajectories from failed rollouts and pair them with successful expert demonstrations to create contrastive training data… uses ‘Policy-Preserving Augmentation’ (PPA) to ensure the model learns robust policy logic rather than just memorizing tool-specific schemas

Hyperight coverage of AED release hyperight.com

50,228 pairs from 9,961 source tasks… released under CC BY-NC-SA 4.0 — any adaptations or derivative works built using the dataset must be redistributed under the same license terms, effectively ‘infecting’ downstream research with the same restrictions

HarnessFix (OpenReview) openreview.net

compiles raw traces into a ‘Harness-aware Trace Intermediate Representation’ to isolate whether a failure stemmed from the model’s reasoning or the ‘agent harness’ (the runtime environment and tool interfaces)… principled debugging of the harness is often more effective than model fine-tuning

Moonlight review of LANTERN themoonlight.io

the final ‘content screen’ that reduced 62 verified relations to 13 non-trivial ones suggests a high rate of technically correct but mathematically uninteresting output… [LANTERN] lacks the rigorous ‘proof-first’ approach of systems like AlphaProof, relying instead on empirical verification

Moonlight review (ES) — memorization critique themoonlight.io

critics from the ‘Where Induction Runs Out’ study argue that OEIS-derived benchmarks often measure memorization rather than algorithmic induction… high performance on OEIS tasks often tracks the model’s recognition of a sequence from its pre-training corpus

MDPI survey on OEIS discovery methods mdpi.com

symbolic approaches like Diofantos… reformulating equation discovery as a Diophantine problem to achieve over 90% reconstruction rates on known OEIS formulas, markedly higher than the ~70% achieved by approximate methods like SINDy

OpenReview — linear probing of latent solvability openreview.net

an LLM’s internal belief about whether a problem is solvable is encoded as a distinct, linearly decodable direction, separate from the direction used for verbalization

arXiv:2609.32264 (abstract) arxiv.org

the Rogers-Ramanujan continued fraction R(q) and a specific theta quotient T(q) can be combined as T(q)^2 R(q)^5 to form the generating function for 5-core partitions

Hacker News discussion news.ycombinator.com

community members have specifically debated the emergence of ‘interpretative meta-tokens’… and how these correlate with the mathematical ‘hidden knowledge’ LANTERN identifies