An automatic numeric-remapping attack generator reveals 12-26 point accuracy drops on GSM8K for three LLMs while MAWPS and MultiArith stay near 98%.
hub Canonical reference
Graph of thoughts: Solving elaborate problems with large language models
Canonical reference. 100% of citing Pith papers cite this work as background.
23
Pith papers citing it
Background
100% of classified citations
hub tools
citation-role summary
background 11
citation-polarity summary
roles
background 11polarities
background 11representative citing papers
Agentic interpretation uses lattices to track LLM judgments on decomposed program claims during analysis.
A new structured prompting method (SPEC) helps AI detect insufficient evidence in adjudication tasks and defer decisions appropriately, reaching 89% accuracy on a benchmark varying information completeness from Colorado unemployment insurance cases.
PTR framework profiles a workflow upfront then executes it deterministically with bounded verification and repair, limiting LM calls to 2-3 while outperforming ReAct in 16 of 24 tested configurations.
Evaluations of 53 LLMs on 14 basic math tasks show reasoning models use ~18x more tokens with sometimes lower accuracy, non-monotonic gains from extended budgets, and sharp performance drops under token constraints.
CODI compresses explicit CoT into continuous space via self-distillation and is the first implicit method to match explicit CoT performance on GSM8k at GPT-2 scale with 3.1x compression and 28.2% higher accuracy than prior implicit approaches.
Bidirectional Evolutionary Search augments autoregressive expansion with evolutionary recombination operators and dense backward subgoal feedback to produce better candidates than standard best-of-N or tree search for language model self-improvement.
LLM-X is a scalable architecture for direct negotiation and communication among personal LLM agents, featuring federated gateways, typed protocols, and policy enforcement, shown stable in experiments with up to 12 agents.
LASAR uses two-stage supervised training plus reinforcement learning to ground semantic IDs, align latent reasoning trajectories to CoT hidden states via KL divergence, and adaptively choose reasoning depth, halving average steps while improving quality on three datasets.
GRIL uses stage-specific RL rewards to train LLMs to detect missing premises, pause proactively, and resume grounded reasoning after clarification, yielding up to 45% better premise detection and 30% higher task success on insufficient math datasets.
FACT-E uses controlled perturbations as an instrumental signal to measure intra-chain faithfulness in CoT reasoning and combines it with answer consistency to select trustworthy trajectories.
SeLaR selectively applies latent soft reasoning in LLMs via entropy gating and contrastive regularization, outperforming standard CoT on five benchmarks without training.
Survey that defines agentic RL for LLMs via POMDPs, introduces a taxonomy of planning/tool-use/memory/reasoning capabilities and domains, and compiles open environments from over 500 papers.
Agent Q integrates MCTS-guided search, self-critique, and off-policy DPO to train LLM agents that outperform behavior cloning and reinforced fine-tuning baselines in WebShop and achieve up to 95.4% success in real-world booking scenarios.
Repeated sampling scales problem coverage log-linearly with sample count, improving SWE-bench Lite performance from 15.9% to 56% using 250 samples.
TraceGraph constructs shared state graphs from multi-model trajectories to expose productive cores and trap regions, then uses them to diagnose navigation differences across benchmarks and to drive a recovery pipeline that improves SWE-bench resolved rates by 3-4 points on fired instances.
RGoT uses RL to adaptively generate task-specific graphs of operations for GoT-style LLM prompting from a human-provided set, with results suggesting feasibility under constraints.
SABA improves LLM performance on detective puzzle benchmarks by recursively fusing information into a base state and using queries to resolve missing premises before concluding.
The paper unifies perspectives on Long CoT in reasoning LLMs by introducing a taxonomy, detailing characteristics of deep reasoning and reflection, and discussing emergence phenomena and future directions.
Outcome optimization induces reward-induced manifold collapse in LLMs by favoring low-complexity spurious correlations over high-complexity causal reasoning, with process reward models acting as topological filters to block shortcuts.
OneSearch-V2 improves generative retrieval via latent reasoning and self-distillation, achieving +3.98% item CTR, +2.07% buyer volume, and +2.11% order volume in online A/B tests.
The paper proposes a bottom-up framework for safe agentic AI systems that treats each component as a dual-use interface where added capabilities also expand attack surfaces across single agents, multi-agent systems, and interoperable ecosystems.
The paper introduces a taxonomy of AI safety for LLMs organized into Trustworthy AI, Responsible AI, and Safe AI perspectives, accompanied by a review of state-of-the-art methods, challenges, and future directions.
citing papers explorer
-
Testing LLM Arithmetic Reasoning Generalization with Automatic Numeric-Remapping Attacks
An automatic numeric-remapping attack generator reveals 12-26 point accuracy drops on GSM8K for three LLMs while MAWPS and MultiArith stay near 98%.
-
Agentic Interpretation: Lattice-Structured Evidence for LLM-Based Program Analysis
Agentic interpretation uses lattices to track LLM judgments on decomposed program claims during analysis.
-
Learning When Not to Decide: A Framework for Overcoming Factual Presumptuousness in AI Adjudication
A new structured prompting method (SPEC) helps AI detect insufficient evidence in adjudication tasks and defer decisions appropriately, reaching 89% accuracy on a benchmark varying information completeness from Colorado unemployment insurance cases.
-
Profile-Then-Reason: Bounded Semantic Complexity for Tool-Augmented Language Agents
PTR framework profiles a workflow upfront then executes it deterministically with bounded verification and repair, limiting LM calls to 2-3 while outperforming ReAct in 16 of 24 tested configurations.
-
Do LLMs Overthink Basic Math Reasoning? Benchmarking the Accuracy-Efficiency Tradeoff in Language Models
Evaluations of 53 LLMs on 14 basic math tasks show reasoning models use ~18x more tokens with sometimes lower accuracy, non-monotonic gains from extended budgets, and sharp performance drops under token constraints.
-
CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillation
CODI compresses explicit CoT into continuous space via self-distillation and is the first implicit method to match explicit CoT performance on GSM8k at GPT-2 scale with 3.1x compression and 28.2% higher accuracy than prior implicit approaches.
-
Self-Improving Language Models with Bidirectional Evolutionary Search
Bidirectional Evolutionary Search augments autoregressive expansion with evolutionary recombination operators and dense backward subgoal feedback to produce better candidates than standard best-of-N or tree search for language model self-improvement.
-
LLM-X: A Scalable Negotiation-Oriented Exchange for Communication Among Personal LLM Agents
LLM-X is a scalable architecture for direct negotiation and communication among personal LLM agents, featuring federated gateways, typed protocols, and policy enforcement, shown stable in experiments with up to 12 agents.
-
LASAR: Latent Adaptive Semantic Aligned Reasoning for Generative Recommendation
LASAR uses two-stage supervised training plus reinforcement learning to ground semantic IDs, align latent reasoning trajectories to CoT hidden states via KL divergence, and adaptively choose reasoning depth, halving average steps while improving quality on three datasets.
-
Pause or Fabricate? Training Language Models for Grounded Reasoning
GRIL uses stage-specific RL rewards to train LLMs to detect missing premises, pause proactively, and resume grounded reasoning after clarification, yielding up to 45% better premise detection and 30% higher task success on insufficient math datasets.
-
FACT-E: Causality-Inspired Evaluation for Trustworthy Chain-of-Thought Reasoning
FACT-E uses controlled perturbations as an instrumental signal to measure intra-chain faithfulness in CoT reasoning and combines it with answer consistency to select trustworthy trajectories.
-
SeLaR: Selective Latent Reasoning in Large Language Models
SeLaR selectively applies latent soft reasoning in LLMs via entropy gating and contrastive regularization, outperforming standard CoT on five benchmarks without training.
-
The Landscape of Agentic Reinforcement Learning for LLMs: A Survey
Survey that defines agentic RL for LLMs via POMDPs, introduces a taxonomy of planning/tool-use/memory/reasoning capabilities and domains, and compiles open environments from over 500 papers.
-
Agent Q: Advanced Reasoning and Learning for Autonomous AI Agents
Agent Q integrates MCTS-guided search, self-critique, and off-policy DPO to train LLM agents that outperform behavior cloning and reinforced fine-tuning baselines in WebShop and achieve up to 95.4% success in real-world booking scenarios.
-
Large Language Monkeys: Scaling Inference Compute with Repeated Sampling
Repeated sampling scales problem coverage log-linearly with sample count, improving SWE-bench Lite performance from 15.9% to 56% using 250 samples.
-
TraceGraph: Shared Decision Landscapes for Diagnosing and Improving Agent Trajectories
TraceGraph constructs shared state graphs from multi-model trajectories to expose productive cores and trap regions, then uses them to diagnose navigation differences across benchmarks and to drive a recovery pipeline that improves SWE-bench resolved rates by 3-4 points on fired instances.
-
Reinforced Graph of Thoughts: RL-Driven Adaptive Prompting for LLMs
RGoT uses RL to adaptively generate task-specific graphs of operations for GoT-style LLM prompting from a human-provided set, with results suggesting feasibility under constraints.
-
Self-Awareness before Action: Mitigating Logical Inertia via Proactive Cognitive Awareness
SABA improves LLM performance on detective puzzle benchmarks by recursively fusing information into a base state and using queries to resolve missing premises before concluding.
-
Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models
The paper unifies perspectives on Long CoT in reasoning LLMs by introducing a taxonomy, detailing characteristics of deep reasoning and reflection, and discussing emergence phenomena and future directions.
-
The Paradox of Outcome Optimization: A Causal Information-Theoretic Bound on Reasoning Shortcuts in LLMs
Outcome optimization induces reward-induced manifold collapse in LLMs by favoring low-complexity spurious correlations over high-complexity causal reasoning, with process reward models acting as topological filters to block shortcuts.
-
OneSearch-V2: The Latent Reasoning Enhanced Self-distillation Generative Search Framework
OneSearch-V2 improves generative retrieval via latent reasoning and self-distillation, achieving +3.98% item CTR, +2.07% buyer volume, and +2.11% order volume in online A/B tests.
-
Toward a Safe Internet of Agents
The paper proposes a bottom-up framework for safe agentic AI systems that treats each component as a dual-use interface where added capabilities also expand attack surfaces across single agents, multi-agent systems, and interoperable ecosystems.
-
AI Safety Landscape for Large Language Models: Taxonomy, State-of-the-art, and Future Directions
The paper introduces a taxonomy of AI safety for LLMs organized into Trustworthy AI, Responsible AI, and Safe AI perspectives, accompanied by a review of state-of-the-art methods, challenges, and future directions.