While humans sometimes do show the capability of correcting their own erroneous guesses with self-critiquing, there seems to be no basis for that assumption in the case of LLMs.
Chain-of-Thought (CoT) prompting can enhance the reasoning capabilities of large language models (LLMs), establishing itself as a primary approach to solving complex reasoning tasks. Existing CoT synthesis approaches usually focus on simpler reasoning tasks and thus result in low-quality and inconsistent CoT prompts. In response to this challenge, we present an empirical investigation of CoT prompting and introduce CoTGenius, a novel framework designed for the automatic generation of superior CoT prompts. CoTGenius is developed based on three major evolution strategies, i.e., complicate, diversify, and specify — alongside two filtering mechanisms: evolutionary success judgement and correctness verification. We further employ CoTGenius to create an extensive CoT dataset, and subsequently fine-tune the Llama 2-Chat 7B and 13B models on this dataset. We call the resulting model ChainLM. To deal with the cumulative error issue in reasoning steps, we propose a step-level debating method, wherein multiple debaters discuss each reasoning step to arrive at the correct answer. Extensive experiments demonstrate that our ChainLM models exhibit enhanced proficiency in addressing a spectrum of complex reasoning problems compared to existing models. In addition, we conduct an in-depth analysis of the impact of data categories within CoTGe
Large Language Models (LLMs) have achieved impressive performance across various reasoning tasks. However, even state-of-the-art LLMs such as ChatGPT are prone to logical errors during their reasoning processes. Traditional approaches to mitigate these errors involve human or tool-based feedback, such as employing task-specific verifiers or aggregating multiple reasoning paths. These methods, however, either depend heavily on human input or struggle with inconsistent responses. To overcome these limitations, we present RankPrompt, an innovative prompting strategy that empowers LLMs to autonomously rank their responses without needing extra resources. RankPrompt simplifies the ranking challenge into comparative evaluations among different responses, leveraging LLMs’ innate ability to generate comparative examples within context. Our experiments across 11 arithmetic and commonsense reasoning tasks show that RankPrompt significantly enhances the reasoning performance of ChatGPT and GPT-4, with improvements of up to 13%. Furthermore, RankPrompt shows exceptional performance in LLM-based automatic evaluations for open-ended tasks, matching human judgments 74% of the time in the AlpacaEval dataset. It also proves to be robust against changes in response order and inconsistency. Overall, our findings endorse RankPrompt as an effective method for extracting high-quality feedback directly from language models.
This paper introduces RankPrompt, a novel prompting method that enhances the reasoning capabilities of large language models (LLMs) like ChatGPT and GPT-4. The key ideas are:
1. Generate multiple diverse reasoning paths (candidates) for a given question using few-shot chain-of-thought prompting.
2. Guide the LLM to comparatively evaluate and rank the candidate reasoning paths in a step-by-step manner, leveraging carefully designed prompts and automatically generated comparison exemplars.
3. Select the top-ranked reasoning path as the final answer.
The main advantages of RankPrompt are: 1) It does not require additional models or human annotations, 2) It achieves strong performance across various reasoning and automatic evaluation tasks, outperforming baseline methods, and 3) It is robust to inconsistent reasoning paths.
Experiments on 11 arithmetic, commonsense, and symbolic reasoning tasks demonstrate RankPrompt’s superiority over baselines like chain-of-thought prompting and majority voting, with accuracy improvements of up to 13%. On the AlpacaEval benchmark for automatic evaluation, RankPrompt achieves 74% agreement with human judgments, setting a new state-of-the-art for LLM-based evaluators.
Modern systems for multi-hop question answering (QA) typically break questions into a sequence of reasoning steps, termed chain-ofthought (CoT), before arriving at a final answer. Often, multiple chains are sampled and aggregated through a voting mechanism over the final answers, but the intermediate steps themselves are discarded. While such approaches improve performance, they do not consider the relations between intermediate steps across chains and do not provide a unified explanation for the predicted answer. We introduce MultiChain Reasoning (MCR), an approach which prompts large language models to meta-reason over multiple chains of thought, rather than aggregate their answers. MCR examines different reasoning chains, mixes information between them and selects the most relevant facts in generating an explanation and predicting the answer. MCR outperforms strong baselines on 7 multi-hop QA datasets. Moreover, our analysis reveals that MCR explanations exhibit high quality, enabling humans to verify its answers.
In this paper, we aim to improve the reasoning ability of large language models (LLMs) over knowledge graphs (KGs) to answer complex questions. Inspired by existing methods that design the interaction strategy between LLMs and KG, we propose an autonomous LLM-based agent framework, called KG-Agent, which enables a small LLM to actively make decisions until finishing the reasoning process over KGs. In KG-Agent, we integrate the LLM, multifunctional toolbox, KG-based executor, and knowledge memory, and develop an iteration mechanism that autonomously selects the tool then updates the memory for reasoning over KG. To guarantee the effectiveness, we leverage program language to formulate the multi-hop reasoning process over the KG, and synthesize a code-based instruction dataset to fine-tune the base LLM. Extensive experiments demonstrate that only using 10K samples for tuning LLaMA-7B can outperform state-of-theart methods using larger LLMs or more data, on both in-domain and out-domain datasets. Our code and data will be publicly released.
In this work, we proposed an autonomous agent framework to synergize LLMs and KGs to perform complex reasoning over KG, namely KG-Agent. In our approach, we first curated a toolbox for KG, consisting of three types of tools to support the typical operations when reasoning on KG. Then, we developed an autonomous iteration mechanism based on tool selection and memory updation that integrates the LLM, multifunctional toolbox, KGbased executor, and knowledge memory, for reasoning over KG. Next, we leveraged existing KGQA datasets to synthesize the code-based instruction tuning dataset. Finally, with only 10K tuning samples, we implemented the autonomous agent relying on the smaller 7B LLM, which mostly outperforms state-of-the-art baselines based on full-data tuning or larger LLMs. In future work, we will consider extending our framework to deal with more types of structured data, e.g., databases and tables.
Foundation models in language and vision have the ability to run inference on any textual and visual inputs thanks to the transferable representations such as a vocabulary of tokens in language. Knowledge graphs (KGs) have different entity and relation vocabularies that generally do not overlap. The key challenge of designing foundation models on KGs is to learn such transferable representations that enable inference on any graph with arbitrary entity and relation vocabularies. In this work, we make a step towards such foundation models and present ULTRA, an approach for learning universal and transferable graph representations. ULTRA builds relational representations as a function conditioned on their interactions. Such a conditioning strategy allows a pre-trained ULTRA model to inductively generalize to any unseen KG with any relation vocabulary and to be fine-tuned on any graph. Conducting link prediction experiments on 57 different KGs, we find that the zero-shot inductive inference performance of a single pre-trained ULTRA model on unseen graphs of various sizes is often on par or better than strong baselines trained on specific graphs. Fine-tuning further boosts the performance.
Creating LLM-generated knowledge graphs
We note the basic flow that underpins GraphRAG, which builds upon our prior research(opens in new tab) and repositories(opens in new tab) using graph machine learning:
The LLM processes the entire private dataset, creating references to all entities and relationships within the source data, which are then used to create an LLM-generated knowledge graph.
This graph is then used to create a bottom-up clustering that organizes the data hierarchically into semantic clusters (indicated by using color in Figure 3 below). This partitioning allows for pre-summarization of semantic concepts and themes, which aids in holistic understanding of the dataset.
At query time, both of these structures are used to provide materials for the LLM context window when answering a question.
Knowledge graph embeddings (KGEs) were originally developed to infer true but missing facts in incomplete knowledge repositories. In this paper, we link knowledge graph completion and counterfactual reasoning via our new task CFKGR. We model the original world state as a knowledge graph, hypothetical scenarios as edges added to the graph, and plausible changes to the graph as inferences from logical rules. We create corresponding benchmark datasets, which contain diverse hypothetical scenarios with plausible changes to the original knowledge graph and facts that should be retained. We develop COULDD, a general method for adapting existing knowledge graph embeddings given a hypothetical premise, and evaluate it on our benchmark. Our results indicate that KGEs learn patterns in the graph without explicit training. We further observe that KGEs adapted with COULDD solidly detect plausible counterfactual changes to the graph that follow these patterns. An evaluation on human-annotated data reveals that KGEs adapted with COULDD are mostly unable to recognize changes to the graph that do not follow learned inference rules. In contrast, ChatGPT mostly outperforms KGEs in detecting plausible changes to the graph but has poor knowledge retention. In summary, CFKGR connects two previously distinct areas, namely KG completion and counterfactual reasoning.
