News without the noise
Benchmark
Reported in 43 papers on Zotpaper. The latest, each as the paper states it:
20 results for “WebShop”
Research · Agents
CORD: Conflict-Aware Optimization for Joint RL and Distillation in Agents
CORD selectively removes opposing gradient components while preserving a protected objective, boosting task success on ALFWorld and WebShop across several training recipes.
WebShop: 64.6% success rate (as reported)
The Chinese University of Hong Kong, Tencent, Peking University
Research · Agents
SAPD: Learning from Agent Experience without Environment Interaction
SAPD decomposes recorded multi-turn histories into decision prefixes and uses relative step advantages to stabilize policy distillation, improving results on ALFWorld, WebShop and Search-based QA.
WebShop: 10.9 x per-step training time speedup vs online OPD (as reported)
Meituan, University of Chinese Academy of Sciences
Research · Reinforcement learning
BLPO: Beyond-Local Policy Optimization for Long-Horizon LLM Agents
Across two simulated environments and two model sizes, the authors report higher task success when training reuses evidence about useful actions across tasks and training updates.
WebShop: 88.9 Task Score (as reported)
Beihang University, Alibaba Group
Research · Agents
TalentRL: Strategy-Guided Reinforcement Learning for Hierarchical Skill Routing
A hierarchical reinforcement learning approach separates strategy selection from skill execution to improve routing credit assignment.
WebShop: 90.7 average score (as reported)
Microsoft AI
Research · Evaluation & benchmarks
IntentDelta: A Human-Constructed Benchmark for Evolving User Intent
A human-constructed benchmark tests eleven agents on shopping and travel tasks, finding weaknesses in both recovering revised intentions and executing them.
WebShop: 44.7% Intention Recovery (as reported)
Georgia Institute of Technology, Microsoft Research
Research · Reinforcement learning
Contextual Credit Policy Optimization
CCPO uses context-weighted trajectory comparisons and short-horizon progress to identify which actions contributed to eventual success.
WebShop: 83.6 mean task score (as reported)
Stanford University, Texas A&M University, Yale University
Research · Agents
VETTA: Coordinating Turn- and Token-Level Credit Assignment for Multi-Turn LLM Agents
VETTA learns both credit types via separate value heads on a lightweight shared critic, improving success rates on two agent benchmarks.
WebShop: 83.9 task score (as reported)
University of Science and Technology of China, Zhongguancun Academy, Shandong University
Research · Reinforcement learning
TrajEvolve: Trajectory Evolution for Reinforcement Learning with Hindsight Credit Assignment
TrajEvolve revises unnecessary steps in historical trajectories and uses the resulting paths alongside fresh policy rollouts during training.
WebShop: 82% success rate (as reported)
City University of Hong Kong, Tencent, Nanjing University
Research · Agents
From Local Skills to Long-Horizon Web Tasks: Progressive Skill Exploration for LLM Web Agents
A training loop that extracts reusable sub-skills from unsuccessful web-browsing trajectories, stitches them into longer skill chains, and distils the results back into the model improved task success by 3.9% to 18.6% over the strongest baselines.
WebShop: 0.93 reward (as reported)
Renmin University of China, Huawei Technologies
Research · Language models
Branching On-Policy Distillation
Chen et al. show that resampling continuations at informative decision points broadens teacher supervision and outperforms independent rollouts with fewer student tokens.
WebShop: 33.59% success rate (as reported)
Tsinghua University, SenseTime Research, Peking University
Research · Reinforcement learning
ProAS: A Progress-Aware Advantage Shaping Framework for Multi-Turn Agentic Reinforcement Learning
ProAS uses a reference graph to mask invalid actions and penalize early deviations, boosting success and convergence.
WebShop: 89.1 task score (as reported)
Zhejiang University, Westlake University, Westlake Institute for Advanced Study
Research · Agents
Learning from Interaction Experience via Localized Context Distillation
The authors propose a method that identifies and revises only critical spans in a student's trajectory, outperforming both off-policy and on-policy context distillation across math, code, and agent tasks.
WebShop: 52.4% task success rate (as reported)
Tencent, The Chinese University of Hong Kong, Zhejiang University
Research · Agents
Sibyl: An Efficient Small-large Model Collaboration Framework for Long-horizon Tasks
Sibyl trains a 0.6-billion-parameter model to request step-level assistance only when useful, then absorb that guidance into later decisions.
WebShop: 91.1 score (as reported)
Hangzhou Dianzi University, Fudan University, The Hong Kong University of Science and Technology
Research · Agents
Procedural Exposure Bias in Skill-Augmented Language Model Agents
The same procedural content, shown as text versus as an image, leads to measurable differences in action preference and downstream agent success.
ALFWorld, ScienceWorld, WebShop: 58.64% macro-average success rate (as reported)
Northeastern University, Alibaba Group
Research · Agents
Beyond the Teacher: Understanding and Preserving Distillation Overshoot in Agentic On-Policy Distillation
Zhong et al. show that on-policy distillation repairs weak student decisions early but erodes successful ones later, and propose SPOD to protect them.
WebShop: 42.8% success (as reported)
University of Southern California, University of Florida, Emory University
Research · Reinforcement learning
From Outcomes to Steps: Reliable Success-Aware Critic Learning for Long-Horizon Language Agents
SCALE adds an explicit success probability head to actor-critic RL and uses reliability-weighted changes in that probability to credit intermediate actions, raising success rates on five agent benchmarks.
WebShop: 83.8% success rate (as reported)
Shanghai Jiao Tong University, Huawei Noah's Ark Lab
Research · Agents
Co-Evolving Acting and Forecasting for Language Agents
CoPE trains agents to predict several subsequent actions from each interaction history, improving five benchmark results without adding inference-time computation.
WebShop: 77% success rate (as reported)
University of Illinois Urbana-Champaign, University of California San Diego, Meta AI
Research · Agents
When Should Agents Think? Adaptive Reasoning via Cross-Turn Estimation
RACE trains language-model agents to estimate how long prior reasoning remains useful, reducing repeated reasoning across four interactive benchmarks.
WebShop: 535 tokens Avg Tok. (as reported)
Renmin University of China, Alibaba Token Hub, Alibaba Group
Research · Agents
State-OPD: State-Grounded On-Policy Distillation for Multi-Turn Agents
State-OPD separates executed interaction history from student reasoning, improving teacher supervision and broadening the actions sampled during training.
WebShop: 19.65% success rate (as reported)
Alibaba Group, Tsinghua University, Shanghai Jiao Tong University
Research · Agents
ASCENT: Online Test-Time Training of Long-Horizon Agents via Self-Distillation of Verified Experience
ASCENT enables online test-time training of LLM agents by self-distilling successful trajectories into persistent weight updates without external supervision.
WebShop: 41.5% strict success rate (as reported)
University of New South Wales (UNSW Sydney)