In 30-step recursive LLM loops, append-mode persistent escape from source basins reaches 50% near 400 tokens under full history but plateaus below 50% under tail-clip memory policy, while replace-mode switching largely reflects state reset.
hub
Generating sequences by learning to self-correct
13 Pith papers cite this work, alongside 30 external citations. Polarity classification is still indexing.
13
Pith papers citing it
30
external citations · external index
hub tools
representative citing papers
VeriGUI adds a Thinking-Verification-Action-Expectation loop and two-stage training on synthetic failures to reduce undetected action errors and improve recovery in GUI automation.
LLMs show strong user bias in role-tagged contexts that is amplified by preference alignment and can be reduced or controlled through targeted fine-tuning and DPO.
HQRE entropy regularization makes multi-agent LLM coordination well-posed, yielding unique equilibria, linear mirror convergence, bounded Bayesian regret, and DICE gains of 4.3–8.5 pp on reasoning/planning tasks.
DRIFT achieves multi-turn RL performance via offline importance-weighted SFT by leveraging the equivalence of KL-regularized RL to weighted supervised learning.
DenoiseRL injects wrong reasoning prefixes from weak models into RL rollouts and trains the policy to recover, improving GRPO/DAPO mathematical reasoning on named benchmarks.
STRIDE co-trains generator and verifier on outcome rewards alone to deliver learnable stepwise language feedback that redirects LLM reasoning trajectories and outperforms scalar-reward baselines.
LiveCodeBench collects 400 recent contest problems to create a contamination-free benchmark evaluating LLMs on code generation and related capabilities like self-repair and execution.
RAP turns LLMs into dual world-model and planning agents via MCTS to generate better reasoning paths, outperforming CoT baselines and achieving 33% relative gains over GPT-4 CoT using LLaMA-33B on plan generation.
Pre-trained LLMs using recursive criticism and improvement prompting achieve state-of-the-art results on the MiniWoB++ computer task benchmark with only a handful of demonstrations and no task-specific reward function.
EPPC-OASIS combines ontology-aware fine-tuning via Wasserstein alignment with structured inference refinement to extract EPPC codes from secure messages, reporting 77.13% Code+Sub-code F1 and 63.83% Triplet F1 with small gains over supervised fine-tuning baselines.
A co-evolving proposer-critic RL framework improves GUI grounding accuracy by letting the model critique its own proposals rendered on screenshots.
Self-Refine boosts LLM outputs by ~20% on average across seven tasks by having the same model iteratively generate, critique, and refine its own responses.
citing papers explorer
-
Perturbation Dose Responses in Recursive LLM Loops: Raw Switching, Stochastic Floors, and Persistent Escape under Append, Replace, and Dialog Updates
In 30-step recursive LLM loops, append-mode persistent escape from source basins reaches 50% near 400 tokens under full history but plateaus below 50% under tail-clip memory policy, while replace-mode switching largely reflects state reset.
-
Don't Act Blindly: Robust GUI Automation via Action-Effect Verification and Self-Correction
VeriGUI adds a Thinking-Verification-Action-Expectation loop and two-stage training on synthetic failures to reduce undetected action errors and improve recovery in GUI automation.
-
User-Assistant Bias in LLMs
LLMs show strong user bias in role-tagged contexts that is amplified by preference alignment and can be reduced or controlled through targeted fine-tuning and DPO.
-
DICE: Entropy-Regularized Equilibrium Selection for Stable Multi-Agent LLM Coordination
HQRE entropy regularization makes multi-agent LLM coordination well-posed, yielding unique equilibria, linear mirror convergence, bounded Bayesian regret, and DICE gains of 4.3–8.5 pp on reasoning/planning tasks.
-
DRIFT: Decoupled Rollouts and Importance-Weighted Fine-Tuning for Efficient Multi-Turn Optimization
DRIFT achieves multi-turn RL performance via offline importance-weighted SFT by leveraging the equivalence of KL-regularized RL to weighted supervised learning.
-
DenoiseRL: Bootstrapping Reasoning Models to Recover from Noisy Prefixes
DenoiseRL injects wrong reasoning prefixes from weak models into RL rollouts and trains the policy to recover, improving GRPO/DAPO mathematical reasoning on named benchmarks.
-
STRIDE: Learnable Stepwise Language Feedback for LLM Reasoning
STRIDE co-trains generator and verifier on outcome rewards alone to deliver learnable stepwise language feedback that redirects LLM reasoning trajectories and outperforms scalar-reward baselines.
-
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
LiveCodeBench collects 400 recent contest problems to create a contamination-free benchmark evaluating LLMs on code generation and related capabilities like self-repair and execution.
-
Reasoning with Language Model is Planning with World Model
RAP turns LLMs into dual world-model and planning agents via MCTS to generate better reasoning paths, outperforming CoT baselines and achieving 33% relative gains over GPT-4 CoT using LLaMA-33B on plan generation.
-
Language Models can Solve Computer Tasks
Pre-trained LLMs using recursive criticism and improvement prompting achieve state-of-the-art results on the MiniWoB++ computer task benchmark with only a handful of demonstrations and no task-specific reward function.
-
EPPC-OASIS: Ontology-Aware Adaptation and Structured Inference Refinement for Electronic Patient-Provider Communication Mining in Secure Messages
EPPC-OASIS combines ontology-aware fine-tuning via Wasserstein alignment with structured inference refinement to extract EPPC codes from secure messages, reporting 77.13% Code+Sub-code F1 and 63.83% Triplet F1 with small gains over supervised fine-tuning baselines.
-
Measure Twice, Click Once: Co-evolving Proposer and Visual Critic via Reinforcement Learning for GUI Grounding
A co-evolving proposer-critic RL framework improves GUI grounding accuracy by letting the model critique its own proposals rendered on screenshots.
-
Self-Refine: Iterative Refinement with Self-Feedback
Self-Refine boosts LLM outputs by ~20% on average across seven tasks by having the same model iteratively generate, critique, and refine its own responses.