REVIEW 44 cited by
O1 Replication Journey: A Strategic Progress Report -- Part 1
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
This paper introduces a pioneering approach to artificial intelligence research, embodied in our O1 Replication Journey. In response to the announcement of OpenAI's groundbreaking O1 model, we embark on a transparent, real-time exploration to replicate its capabilities while reimagining the process of conducting and communicating AI research. Our methodology addresses critical challenges in modern AI research, including the insularity of prolonged team-based projects, delayed information sharing, and the lack of recognition for diverse contributions. By providing comprehensive, real-time documentation of our replication efforts, including both successes and failures, we aim to foster open science, accelerate collective advancement, and lay the groundwork for AI-driven scientific discovery. Our research progress report diverges significantly from traditional research papers, offering continuous updates, full process transparency, and active community engagement throughout the research journey. Technologically, we proposed the journey learning paradigm, which encourages models to learn not just shortcuts, but the complete exploration process, including trial and error, reflection, and backtracking. With only 327 training samples and without any additional tricks, journey learning outperformed conventional supervised learning by over 8\% on the MATH dataset, demonstrating its extremely powerful potential. We believe this to be the most crucial component of O1 technology that we have successfully decoded. We share valuable resources including technical hypotheses and insights, cognitive exploration maps, custom-developed tools, etc at https://github.com/GAIR-NLP/O1-Journey.
Forward citations
Cited by 44 Pith papers
-
ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling
Per-step retrieval of solved exemplars injected into the reasoning trace improves test-time scaling accuracy, with up to 13.4 absolute points gained on AIME 2025.
-
Answer-Conditioned Chains of Thought Degrade Verifiable-Reasoning Distillation in Large Language Models
Answer-conditioned chains of thought are harmful supervision for reasoning distillation—causing up to ~27-point accuracy drops—and the damage is invisible to correctness filters.
-
LAMDAS: LLM as an Implicit Classifier for Domain-specific Data Selection
LAMDAS selects domain-relevant training data via an LLM likelihood ratio with a learned domain prefix, beating full-data training and nine baselines on code and math.
-
Unlocking Recursive Thinking of LLMs: Alignment via Refinement
An offline alignment pipeline using reward-filtered self-refinement data and long chain-of-thought SFT raises an 8B model's AlpacaEval 2 win rate from 25.0% to 51.0% with roughly 14k training examples.
-
GThinker: Towards General Multimodal Reasoning via Cue-Guided Rethinking
A cue-tagging and rethinking training recipe lifts a 7B multimodal model to 81.5% on M3CoT, though the gain may reflect training on the same benchmark.
-
Discriminative Policy Optimization for Token-Level Reward Models
Q-RM learns token-level rewards from preference data via a discriminative policy and improves RL alignment on math, reading, and instruction-following benchmarks.
-
TAT-R1: Terminology-Aware Translation with Reinforcement Learning and Word Alignment
Word-alignment rewards for RL-trained translation raise terminology accuracy on RTT from 54.42 to 56.42 TA without hurting general translation quality.
-
Which Data Attributes Stimulate Math and Code Reasoning? An Investigation via Influence Functions
Influence-function attribution shows high-difficulty math and low-difficulty code training data best improve math and code reasoning, and difficulty-based reweighting improves benchmark performance.
-
Tool-Star: Empowering LLM-Brained Multi-Tool Reasoner via Reinforcement Learning
Tool-Star combines cold-start supervised fine-tuning with a multi-tool self-critic reinforcement learning algorithm and hierarchical rewards to improve LLM tool-use reasoning.
-
THOR-MoE: Hierarchical Task-Guided and Context-Responsive Routing for Neural Machine Translation
A hierarchical routing method that combines predicted task labels with context-aware token routing improves BLEU and reduces activated experts in translation MoE models.
-
DiagnosisArena: Benchmarking Diagnostic Reasoning for Large Language Models
DiagnosisArena, a 1,113-case benchmark from top journals, shows state-of-the-art LLMs achieve at most 51% top-1 diagnostic accuracy, far below clinical-level competence.
-
Parallel Scaling Law for Language Models
Running P parallel streams of a language model with learned prefixes gives performance comparable to multiplying parameters by about k log P + 1, with k around 0.33 to 0.39.
-
Scalable Chain of Thoughts via Elastic Reasoning
Training reasoning models with a fixed two-part token budget, thinking plus solution, makes them robust to truncated thinking and more concise overall.
-
Typhoon T1: An Open Thai Reasoning Model
Structured long-thinking SFT turns a 3B Thai instruct model into a reasoning model that improves on several English benchmarks and can think in Thai, with a fully open recipe.
-
Agentic Reasoning: A Streamlined Framework for Enhancing LLM Reasoning with Agentic Tools
An agentic reasoning framework that adds web search, code execution, and knowledge-graph memory to an LLM, boosting performance on expert-level benchmarks close to proprietary deep research systems.
-
DRT: Deep Reasoning Translation via Long Chain-of-Thought
Training LLMs on long chain-of-thought traces improves English-to-Chinese translation of simile and metaphor sentences from literature compared with standard fine-tuning.
-
Reasoning Through Execution: Unifying Process and Outcome Rewards for Code Generation
A tree-based, inference-only framework combining execution metrics with LLM self-critique improves code generation correctness and efficiency across models and benchmarks.
-
Process-Supervised Reward Models for Verifying Clinical Note Generation: A Scalable Approach Guided by Domain Expertise
A step-level reward model trained on expert-designed synthetic clinical errors detects injected note errors with 98.8% accuracy and selects physician-preferred notes with 56.2% accuracy.
-
Beyond Examples: High-level Automated Reasoning Paradigm in In-Context Learning via MCTS
HiAR-ICL builds reusable 'thought card' reasoning templates via MCTS and shows a 7B model with these templates outperforms GPT-4o on MATH and AMC.
-
BlendServe: Optimizing Offline Inference for Auto-regressive Large Models with Resource-aware Batching
BlendServe combines resource-aware batching with prefix sharing using a resource-aware prefix tree and dual scanner, achieving up to 1.44x throughput vs vLLM/SGLang in offline LLM inference.
-
Cognitive Duality for Adaptive Web Agents
A web agent called CogniWeb reportedly reaches 43.96% success on WebArena with 75% fewer tokens by switching between fast and slow reasoning, but the supplied body is a different paper and the claims could not be checked.
-
Decoupling Knowledge and Reasoning in LLMs: An Exploration Using Cognitive Dual-System Theory
The accuracy gap between fast and slow thinking is proposed as a measure of reasoning contribution, and is used to show that reasoning is domain-specific, scaling mainly reduces overthinking, and knowledge and reasoni...
-
Beyond Isolated Capabilities: Bridging Long CoT Reasoning and Long-Context Understanding
DeepSeek-R1-distilled models show higher multi-document QA accuracy than their base counterparts and flatter position-bias curves, especially with 50-80 documents.
-
How to Train a Leader: Hierarchical Reasoning in Multi-Agent LLMs
A leader LLM trained with a GRPO variant that conditions on frozen agent responses improves both collaborative and zero-shot accuracy on BBH, MATH, and MMLU.
-
Confucius3-Math: A Lightweight High-Performance Reasoning LLM for Chinese K-12 Mathematics Learning
A 14B RL-trained model reports 96.2% on an internal Chinese K-12 benchmark, about 15x lower serving cost than DeepSeek-R1, with three new training tricks.
-
Deciphering Trajectory-Aided LLM Reasoning: An Optimization Perspective
Reasoning trajectories are formalized as pseudo-gradient descent on LLM parameters, making LLM reasoning training a MAML-style meta-learning problem.
-
ManuSearch: Democratizing Deep Search in Large Language Models with a Transparent and Open Multi-Agent Framework
ManuSearch's three-agent design brings open-source LLMs to parity or better with proprietary web-search agents on three reasoning benchmarks, and ORION adds 310 long-tail EN/ZH questions for testing them.
-
Sparks of Tabular Reasoning via Text2SQL Reinforcement Learning
Training LLMs on Text-to-SQL with chain-of-thought supervision and GRPO reinforcement learning is reported to improve zero-shot accuracy on tabular question answering, though the gains are measured by an LLM judge rat...
-
On the Emergence of Thinking in LLMs I: Searching for the Right Intuition
A post-training framework called RLSP, decoupling exploration from correctness rewards, induces backtracking and self-correction in LLMs and improves math reasoning performance.
-
Generating Symbolic World Models via Test-time Scaling of Large Language Models
Best-of-N sampling plus iterative self-critique, called iVML, lets a 7B open LLM generate PDDL planning domains with over 85% and 71% success on two benchmarks, outperforming o1-mini.
-
Step Back to Leap Forward: Self-Backtracking for Boosting Reasoning of Language Models
Training LLMs to append a <backtrack> token after erroneous reasoning steps and to use that signal at inference time improves Countdown accuracy by more than 40 percent over supervised fine-tuning.
-
Reasoning Language Models: A Blueprint
A modular blueprint and open-source framework (x1) that presents existing reasoning language model designs as special cases of one unified toolbox.
-
Don't Command, Cultivate: An Exploratory Study of System-2 Alignment
Encouraging LLMs to analyze user requests step-by-step (System-2 Alignment) modestly improves safety on open-source models, but with trade-offs and limited evidence.
-
O1 Replication Journey -- Part 2: Surpassing O1-preview through Simple Distillation, Big Progress or Bitter Lesson?
Distillation from OpenAI's O1 plus supervised fine-tuning gives 13/30 on AIME2024 versus O1-preview's 12/30, and the paper proposes a transparency index for replication claims.
-
Do LLMs Really Think Step-by-step In Implicit Reasoning?
A prompted LLM showed little linear-probing evidence of computing intermediate arithmetic steps during silent reasoning, whereas a model trained to internalize chain-of-thought did, and both degraded sharply under for...
-
A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Models
A structured literature survey concluding that reasoning capabilities do not automatically make LLMs more trustworthy and can introduce new vulnerabilities in safety, robustness, and privacy.
-
A Survey on Model Extraction Attacks and Defenses for Large Language Models
A taxonomy of model extraction attacks and defenses for large language models, with proposed evaluation metrics and future research directions.
-
Beyond Distillation: Pushing the Limits of Medical LLM Reasoning with Minimalist Rule-Based RL
Medical QA accuracy improves substantially through reinforcement learning with a binary correct-answer reward alone, without supervised fine-tuning on distilled reasoning traces.
-
Long-Short Chain-of-Thought Mixture Supervised Fine-Tuning Eliciting Efficient Reasoning in Large Language Models
Mixing long and short chain-of-thought data with a balanced thinking prompt reduces response length, but it does not reliably improve accuracy over the long-only baseline.
-
Quantitative Analysis of Performance Drop in DeepSeek Model Quantization
A hand-tuned 3-bit quantization scheme, DQ3_K_M, matches 4-bit accuracy on DeepSeek-R1 and V3 while shrinking the models enough for single-machine deployment.
-
Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling
Small LLMs with compute-optimal test-time scaling can outperform much larger models on math benchmarks, but the reported strategy is selected on the same test sets used for evaluation.
-
Generative AI Act II: Test Time Scaling Drives Cognition Engineering
Test-time scaling techniques such as long chain-of-thought, tree search, and self-correction define the paper's 'cognition engineering' paradigm, which it surveys, taxonomizes, and tutorials.
-
Dynamic Chain-of-Thought: Towards Adaptive Deep Reasoning
D-CoT claims to cut reasoning time, steps, and tokens versus DeepSeek R1, but the comparison is between a Python simulation and a real LLM, without any accuracy check.
-
The Hitchhiker's Guide to Agentic AI: From Foundations to Systems
A survey-style reference book mapping the full agentic-AI stack from transformer internals to production deployment, with no new research result.
Discussion (0). Continue with ORCID to comment.