SOTAVerified

Logical Reasoning

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 101–150 of 747 papers

TitleStatusHype
Reversal Blessing: Thinking Backward May Outpace Thinking Forward in Multi-choice Questions—0
Autoregressive Image Generation Guided by Chains of Thought—0
Quantifying Logical Consistency in Transformers via Query-Key Alignment—0
AutoLogi: Automated Generation of Logic Puzzles for Evaluating Reasoning Abilities of Large Language ModelsCode1
Logic Haystacks: Probing LLMs Long-Context Logical Reasoning (Without Easily Identifiable Unrelated Padding)—0
R1-Onevision：An Open-Source Multimodal Large Language Model Capable of Deep ReasoningCode4
Intermediate Languages Matter: Formal Choice Drives Neurosymbolic LLM Reasoning—0
From System 1 to System 2: A Survey of Reasoning Large Language ModelsCode5
Empowering LLMs with Logical Reasoning: A Comprehensive Survey—0
Identifying Features that Shape Perceived Consciousness in Large Language Model-based AI: A Quantitative Study of Human Responses—0
On the logical skills of large language models: evaluations using arbitrarily complex first-order logic problemsCode0
Triangulating LLM Progress through Benchmarks, Games, and Cognitive Tests—0
A Mousetrap: Fooling Large Reasoning Models for Jailbreak with Chain of Iterative Chaos—0
SPPD: Self-training with Process Preference Learning Using Dynamic Value Margin—0
HopRAG: Multi-Hop Reasoning for Logic-Aware Retrieval-Augmented Generation—0
Inference-Time Computations for LLM Reasoning and Planning: A Benchmark and Insights—0
Integrating Expert Knowledge into Logical Programs via LLMsCode0
Unveiling the Magic of Code Reasoning through Hypothesis Decomposition and AmendmentCode2
Beyond Single-Task: Robust Multi-Task Length Generalization for LLMs—0
Exposing Numeracy Gaps: A Benchmark to Evaluate Fundamental Numerical Abilities in Large Language ModelsCode1
Quantifying the Capability Boundary of DeepSeek Models: An Application-Driven Performance Analysis—0
Dialogue-based Explanations for Logical Reasoning using Structured Argumentation—0
The Multilingual Mind : A Survey of Multilingual Reasoning in Language Models—0
Logical Reasoning in Large Language Models: A Survey—0
Logical Lease Litigation: Prolog and LLMs for Rental Law Compliance in New York—0
Logical forms complement probability in understanding language model (and human) performance—0
DMWM: Dual-Mind World Model with Long-Term Imagination—0
Large Language Models Meet Symbolic Provers for Logical Reasoning EvaluationCode1
Structural Reformation of Large Language Model Neuron Encapsulation for Divergent Information Aggregation—0
S^2-MAD: Breaking the Token Barrier to Enhance Multi-Agent Debate Efficiency—0
SymAgent: A Neural-Symbolic Self-Learning Agent Framework for Complex Reasoning over Knowledge Graphs—0
Automating Mathematical Proof Generation Using Large Language Model Agents and Knowledge Graphs—0
Standard Neural Computation Alone Is Insufficient for Logical Intelligence—0
ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoning—0
Enhancing Large Language Model Efficiencyvia Symbolic Compression: A Formal Approach Towards Interpretability—0
Town Hall Debate Prompting: Enhancing Logical Reasoning in LLMs through Multi-Persona Interaction—0
Instantiation-based Formalization of Logical Reasoning Tasks using Language Models and Logical Solvers—0
DBRouting: Routing End User Queries to Databases for Answerability—0
SedarEval: Automated Evaluation using Self-Adaptive RubricsCode0
A Causality-aware Paradigm for Evaluating Creativity of Multimodal Large Language Models—0
JustLogic: A Comprehensive Benchmark for Evaluating Deductive Reasoning in Large Language ModelsCode0
VERUS-LM: a Versatile Framework for Combining LLMs with Symbolic Reasoning—0
PIKE-RAG: sPecIalized KnowledgE and Rationale Augmented GenerationCode7
Assessing the Alignment of FOL Closeness Metrics with Human JudgementCode0
LeapVAD: A Leap in Autonomous Driving via Cognitive Perception and Dual-Process ThinkingCode2
Reasoning with Graphs: Structuring Implicit Knowledge to Enhance LLMs Reasoning—0
TimeLogic: A Temporal Logic Benchmark for Video QA—0
Neural Probabilistic Circuits: Enabling Compositional and Interpretable Predictions through Logical Reasoning—0
Multimodal-to-Text Prompt Engineering in Large Language Models Using Feature Embeddings for GNSS Interference Characterization—0
Enhancing Transformers for Generalizable First-Order Logical Entailment—0
Show:102550
← PrevPage 3 of 15Next →
All datasetsLingOlyBIG-bench (Formal Fallacies Syllogisms Negation)BIG-bench (Penguins In A Table)BIG-bench (Reasoning About Colored Objects)BIG-bench (Temporal Sequences)BIG-bench (Logic Grid Puzzle)BIG-bench (StrategyQA)RuWorldTreeWinograd AutomaticBIG-bench (Logical Fallacy Detection)

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1Claude OpusDelta_NoContext28.8—Unverified
2GPT-4oDelta_NoContext25.1—Unverified
3Gemini 1.5 ProDelta_NoContext23.4—Unverified
4GPT-4Delta_NoContext21.5—Unverified
5Command R+Delta_NoContext11.6—Unverified
6GPT-3.5Delta_NoContext11.2—Unverified
7Mixtral 8x7BDelta_NoContext6.4—Unverified
8Llama 3 8BDelta_NoContext4.9—Unverified
9Llama 3 70BDelta_NoContext2.9—Unverified
10Gemma 7BDelta_NoContext2.2—Unverified
#ModelMetricClaimedVerifiedStatus
1PaLM 2 (few-shot, k=3, Direct)Accuracy64.8—Unverified
2PaLM 2 (few-shot, k=3, CoT)Accuracy57.2—Unverified
3OPT 66B (few-shot, k=3)Accuracy54—Unverified
4PaLM 540B (few-shot, k=3)Accuracy53.6—Unverified
5GPT-NeoX 20B (few-shot, k=3)Accuracy52.8—Unverified
6BLOOM 176B (few-shot, k=3)Accuracy52.8—Unverified
7Chinchilla-70B (few-shot, k=5)Accuracy52.1—Unverified
8Bloomberg GPT 50B (few-shot, k=3)Accuracy50.8—Unverified
9Gopher-280B (few-shot, k=5)Accuracy50.7—Unverified
#ModelMetricClaimedVerifiedStatus
1PaLM 2 (few-shot, k=3, CoT)Accuracy84.9—Unverified
2PaLM 2 (few-shot, k=3, Direct)Accuracy65.8—Unverified
3Chinchilla-70B (few-shot, k=5)Accuracy48.7—Unverified
4PaLM 540B (few-shot, k=3)Accuracy44.5—Unverified
5Gopher-280B (few-shot, k=5)Accuracy40.6—Unverified
6BLOOM 176B (few-shot, k=3)Accuracy40.41—Unverified
7Bloomberg GPT (few-shot, k=3)Accuracy37.67—Unverified
8GPT-NeoX (few-shot, k=3)Accuracy33.56—Unverified
9OPT 66B (few-shot, k=3)Accuracy28.08—Unverified
#ModelMetricClaimedVerifiedStatus
1PaLM 2 (few-shot, k=3, CoT)Accuracy91.2—Unverified
2PaLM 2 (few-shot, k=3, Direct)Accuracy61.2—Unverified
3Chinchilla-70B (few-shot, k=5)Accuracy59.7—Unverified
4Gopher-280B (few-shot, k=5)Accuracy49.2—Unverified
5PaLM 540B (few-shot, k=3)Accuracy38—Unverified
6BLOOM 176B (few-shot, k=3)Accuracy36.8—Unverified
7Bloomberg GPT (few-shot, k=3)Accuracy34.8—Unverified
8OPT 66B (few-shot, k=3)Accuracy31.2—Unverified
9GPT-NeoX (few-shot, k=3)Accuracy26—Unverified
#ModelMetricClaimedVerifiedStatus
1PaLM 2 (few-shot, k=3, CoT)Accuracy100—Unverified
2PaLM 2 (few-shot, k=3, Direct)Accuracy96.4—Unverified
3PaLM 540B (few-shot, k=3)Accuracy39.6—Unverified
4BLOOM 176B (few-shot, k=3)Accuracy36.8—Unverified
5Chinchilla-70B (few-shot, k=5)Accuracy32—Unverified
6Bloomberg GPT (few-shot, k=3)Accuracy29.2—Unverified
7OPT 66B (few-shot, k=3)Accuracy23.6—Unverified
8GPT-NeoX (few-shot, k=3)Accuracy21.2—Unverified
9Gopher-280B (few-shot, k=5)Accuracy19—Unverified
#ModelMetricClaimedVerifiedStatus
1Chinchilla-70B (few-shot, k=5)Accuracy44—Unverified
2PaLM-540B (few-shot, k=5)Accuracy42.4—Unverified
3PaLM-62B (few-shot, k=5)Accuracy36.5—Unverified
4Gopher-280B (few-shot, k=5)Accuracy35.1—Unverified
#ModelMetricClaimedVerifiedStatus
1PaLM-540B (few-shot, k=5)Accuracy73.9—Unverified
2Chinchilla-70B (few-shot, k=5)Accuracy68.3—Unverified
3PaLM-62B (few-shot, k=5)Accuracy65.4—Unverified
4Gopher-280B (few-shot, k=5)Accuracy61—Unverified
#ModelMetricClaimedVerifiedStatus
1Human benchmarkAccuracy 83.7—Unverified
2RuGPT-3 LargeAccuracy 40.7—Unverified
3RuGPT-3 MediumAccuracy 38—Unverified
4RuGPT-3 SmallAccuracy 34—Unverified
#ModelMetricClaimedVerifiedStatus
1Human benchmarkAccuracy87—Unverified
2RuGPT-3 SmallAccuracy57.9—Unverified
3RuGPT-3 MediumAccuracy57.2—Unverified
4RuGPT-3 LargeAccuracy55.5—Unverified
#ModelMetricClaimedVerifiedStatus
1Chinchilla-70B (few-shot, k=5)Accuracy72.1—Unverified
2Gopher-280B (few-shot, k=5)Accuracy58.9—Unverified