StarCraft II is a challenging benchmark for AI agents due to the necessity of both precise micro-level operations and strategic macro-awareness. Previous works, such as Alphastar and SCC, achieve impressive performance on tackling StarCraft II , however, still exhibit deficiencies in long-term strategic planning and strategy interpretability. Emerging large language model (LLM) agents, such as Voyage and MetaGPT, presents the immense potential in solving intricate tasks. Motivated by this, we aim to validate the capabilities of LLMs on StarCraft II, a highly complex RTS game. To conveniently take full advantage of LLMs’ reasoning abilities, we first develop textual StratCraft II environment, called TextStarCraft II, which LLM agent can interact. Secondly, we propose a Chain of Summarization method, including single-frame summarization for processing raw observations and multi-frame summarization for analyzing game information, providing command recommendations, and generating strategic decisions. Our experiment consists of two parts: first, an evaluation by human experts, which includes assessing the LLMs’ mastery of StarCraft II knowledge and the performance of LLM agents in the game; second, the in-game performance of LLM agents, encompassing aspects like win rate and the impact of Chain of Summarization. Experiment results demonstrate that: 1. LLMs possess the relevant knowledge and complex planning abilities needed to address StarCraft II scenarios; 2. Human experts consider the performance of LLM agents to be close to that of an average player who has played StarCraft II for eight years; 3. LLM agents are capable of defeating the built-in AI at the Harder(Lv5) difficulty level. We have open-sourced the code2 and released demo videos3 of LLM agent playing StarCraft II.
Summarization is a core function for limited context systems.
Selecting the “right” amount of information to include in a summary is a difficult task. A good summary should be detailed and entity-centric without being overly dense and hard to follow. To better understand this tradeoff, we solicit increasingly dense GPT-4 summaries with what we refer to as a “Chain of Density” (CoD) prompt. Specifically, GPT-4 generates an initial entitysparse summary before iteratively incorporating missing salient entities without increasing the length. Summaries generated by CoD are more abstractive, exhibit more fusion, and have less of a lead bias than GPT-4 summaries generated by a vanilla prompt. We conduct a human preference study on 100 CNN DailyMail articles and find that that humans prefer GPT-4 summaries that are more dense than those generated by a vanilla prompt and almost as dense as human written summaries. Qualitative analysis supports the notion that there exists a tradeoff between informativeness and readability. 500 annotated CoD summaries, as well as an extra 5,000 unannotated summaries, are freely available on HuggingFace1 .
Answering complex natural language questions often necessitates multi-step reasoning and integrating external information. Several systems have combined knowledge retrieval with a large language model (LLM) to answer such questions. These systems, however, suffer from various failure cases, and we cannot directly train them end-to-end to fix such failures, as interaction with external knowledge is non-differentiable. To address these deficiencies, we define a ReAct-style LLM agent with the ability to reason and act upon external knowledge. We further refine the agent through a ReST-like method that iteratively trains on previous trajectories, employing growing-batch reinforcement learning with AI feedback for continuous self-improvement and self-distillation. Starting from a prompted large model and after just two iterations of the algorithm, we can produce a fine-tuned small model that achieves comparable performance on challenging compositional question-answering benchmarks with two orders of magnitude fewer parameters.
Multi-step reasoning and integrating external information need improvement
Large language models (LLMs) have opened up new possibilities for intelligent agents, endowing them with humanlike thinking and cognitive abilities. In this work, we delve into the potential of large language models (LLMs) in autonomous driving (AD). We introduce DriveMLM, an LLM-based AD framework that can perform close-loop autonomous driving in realistic simulators. To this end, (1) we bridge the gap between the language decisions and the vehicle control commands by standardizing the decision states according to the off-the-shelf motion planning module. (2) We employ a multi-modal LLM (MLLM) to model the behavior planning module of a module AD system, which uses driving rules, user commands, and inputs from various sensors (e.g., camera, lidar) as input and makes driving decisions and provide explanations; This model can plug-andplay in existing AD systems such as Apollo for close-loop driving. (3) We design an effective data engine to collect a dataset that includes decision state and corresponding explanation annotation for model training and evaluation. We conduct extensive experiments and show that our model achieves 76.1 driving score on the CARLA Town05 Long, and surpasses the Apollo baseline by 4.7 points under the same settings, demonstrating the effectiveness of our model. We hope this work can serve as a baseline for autonomous driving with LLMs.
Having the ability to empathize is crucial for accurately representing human behavior during conversations. Despite numerous research aim to improve the cognitive capability of models by incorporating external knowledge, there has been limited attention on the sensible and rational expression of the conversation itself, which are crucial components of the cognitive empathy. Guided by self-presentation theory in sociology, we have designed an innovative categorical approach that segregates historical dialogues into sensible and rational sentences and subsequently elucidate the context through the designed attention mechanism. However, the rational information within the conversation is restricted and the external knowledge used in previous methods have limitations of semantic contradiction and narrow vision field. Considering the impressive performance of LLM in the domain of intelligent agent. We employ LLaMA2–70b as a rational brain to analyze the profound logical information maintained in conversations, which assists the model assessing the balance of sensibility and rationality to produce quality empathetic responses. Experimental evaluations demonstrate that our method outperforms other comparable methods on both automatic and human evaluations
// sensible and rational, “empathetic response generation by introducing emotional inference data and utilizing self-presentation theory to enhance the semantic representation of history context”
