We introduces Crimson, a system that enhances the strategic reasoning capabilities of Large Language Models (LLMs) within the realm of cybersecurity. By correlating CVEs with MITRE ATT&CK techniques, Crimson advances threat anticipation and strategic defense efforts. Our approach includes defining and evaluating cybersecurity strategic tasks, alongside implementing a comprehensive human-in-the-loop data-synthetic workflow to develop the CVE-to-ATT&CK Mapping (CVEM) dataset. We further enhance LLMs' reasoning abilities through a novel Retrieval-Aware Training (RAT) process and its refined iteration, RAT-R. Our findings demonstrate that an LLM fine-tuned with our techniques, possessing 7 billion parameters, approaches the performance level of GPT-4, showing markedly lower rates of hallucination and errors, and surpassing other models in strategic reasoning tasks. Moreover, domain-specific fine-tuning of embedding models significantly improves performance within cybersecurity contexts, underscoring the efficacy of our methodology. By leveraging Crimson to convert raw vulnerability data into structured and actionable insights, we bolster proactive cybersecurity defenses.

Logs produced by extensive software systems are integral to monitoring system behaviors. Advanced log analysis facilitates the detection, alerting, and diagnosis of system faults. Log parsing, which entails transforming raw log messages into structured templates, constitutes a critical phase in the automation of log analytics. Existing log parsers fail to identify the correct templates due to reliance on human-made rules. Besides, These methods focus on statistical features while ignoring semantic information in log messages. To address these challenges, we introduce a cutting-edge \textbf{L}og parsing framework with \textbf{E}ntropy sampling and Chain-of-Thought \textbf{M}erging (Lemur). Specifically, to discard the tedious manual rules. We propose a novel sampling method inspired by information entropy, which efficiently clusters typical logs. Furthermore, to enhance the merging of log templates, we design a chain-of-thought method for large language models (LLMs). LLMs exhibit exceptional semantic comprehension, deftly distinguishing between parameters and invariant tokens. We have conducted experiments on large-scale public datasets. Extensive evaluation demonstrates that Lemur achieves the state-of-the-art performance and impressive efficiency.

Last year Stack Overflow became one of the first websites to announce it would charge AI giants for access to content used to train chatbots. Now the popular Q&A service for coders has signed up its first customer--Google--in what CEO Prashanth Chandrasekar says is the start of a "meaningful" new stream of revenue. The deal is significant, because it remains unclear how broadly Google and other AI developers will pay for content needed for AI projects. Millions of books and websites have fueled the development of AI systems, but most publishers have not been compensated, and some are suing over what they allege is misuse. Many publishers, including Stack Overflow, appear threatened by ChatGPT and other generative AI products, which can answer queries that would have previously sent coders their way.

Earlier this month, Google released its long-awaited system "Gemini," giving users access to its AI image-generation technology for the first time. While most early users agreed that the system was impressive, creating detailed images for text prompts in seconds, users soon discovered that it was difficult to get the system to generate images of white people, and soon viral tweets displayed head-scratching examples such as racially diverse Nazis. Some people faulted Gemini for being "too woke," using Gemini as the latest weapon in an escalating culture war on the importance of recognizing the effects of historical discrimination. Many said it reflected a malaise inside Google, and some criticized the field of "AI ethics" as an embarrassment. The idea that ethical AI work is to blame is wrong.

Following the launch of Copilot Pro in January, Microsoft will start rolling out a slew of upgrades today to make its AI assistant more useful in Windows 11. The biggest addition is support for plugins, something that only worked with Copilot on the web previously. You could, for example, connect Copilot to OpenTable so you can make a reservation from within its chat window. If you're low on groceries, you could start an order using the Instacart plugin. Microsoft says that plugins from Kayak, Klarna and Shopify will also be arriving over the next month.

The Intercept, Raw Story and AlterNet filed separate lawsuits accusing ChatGPT of reproducing news content "verbatim or nearly verbatim" while stripping out important attribution like the author's name. OpenAI asked a court to dismiss that claim, saying the NYT took advantage of a ChatGPT bug that made it recite articles word for word.

Recent work exploring the capabilities of pre-trained large language models (LLMs) has demonstrated their ability to act as general pattern machines by completing complex token sequences representing a wide array of tasks, including time-series prediction and robot control. This paper investigates whether the pattern recognition and sequence modeling capabilities of LLMs can be extended to the domain of knowledge tracing, a critical component in the development of intelligent tutoring systems (ITSs) that tailor educational experiences by predicting learner performance over time. In an empirical evaluation across multiple real-world datasets, we compare two approaches to using LLMs for this task, zero-shot prompting and model fine-tuning, with existing, non-LLM approaches to knowledge tracing. While LLM-based approaches do not achieve state-of-the-art performance, fine-tuned LLMs surpass the performance of naive baseline models and perform on par with standard Bayesian Knowledge Tracing approaches across multiple metrics. These findings suggest that the pattern recognition capabilities of LLMs can be used to model complex learning trajectories, opening a novel avenue for applying LLMs to educational contexts. The paper concludes with a discussion of the implications of these findings for future research, suggesting that further refinements and a deeper understanding of LLMs' predictive mechanisms could lead to enhanced performance in knowledge tracing tasks.

Large Language Models (LLMs) can justify or critique their predictions through discussions with other models or humans, thereby enriching their intrinsic understanding of instances. While proactive discussions in the inference phase have been shown to boost performance, such interactions have not been extensively explored during the training phase. We hypothesize that incorporating interactive discussions into the training process can enhance the models' understanding and improve their reasoning and verbal expression abilities during inference. This work introduces the SAIE framework, which facilitates supportive and adversarial discussions between learner and partner models. The learner model receives responses from the partner, and its parameters are then updated based on this discussion. This dynamic adjustment process continues throughout the training phase, responding to the evolving outputs of the learner model. Our empirical evaluation across various tasks, including math problems, commonsense reasoning, and multi-domain knowledge, demonstrates that models fine-tuned with the SAIE framework outperform those trained with conventional fine-tuning approaches. Furthermore, our method enhances the models' reasoning capabilities, improving both individual and multi-agent inference performance.

Large Language Models (LLM) have dramatically surged in popularity over the recent years. Since the release of ChatGPT, LLMs - especially those with an accessible chat interface - have not only been used by experts, but are also becoming an increasingly common tool with significant benefits for laypeople. To that end, many commercial actors have already begun implementing LLMs in their operations, ranging from customer-facing chatbots to internal decision support systems [14, 6]. The fairness of AI algorithms, including LLMs, has been a pernicious issue, motivating a growing literature and community of AI ethics research [8]. Disparities across gender and race, among other attributes, have especially preoccupied this field [4], leading to efforts to include bias auditing as an important component of AI harm mitigation in policy discussions and regulatory frameworks [28]. Mitigating biases arising from the explicit use of race or gender in the prompt is comparatively straightforward.

Entity abstract summarization aims to generate a coherent description of a given entity based on a set of relevant Internet documents. Pretrained language models (PLMs) have achieved significant success in this task, but they may suffer from hallucinations, i.e. generating non-factual information about the entity. To address this issue, we decompose the summary into two components: Facts that represent the factual information about the given entity, which PLMs are prone to fabricate; and Template that comprises generic content with designated slots for facts, which PLMs can generate competently. Based on the facts-template decomposition, we propose SlotSum, an explainable framework for entity abstract summarization. SlotSum first creates the template and then predicts the fact for each template slot based on the input documents. Benefiting from our facts-template decomposition, SlotSum can easily locate errors and further rectify hallucinated predictions with external knowledge. We construct a new dataset WikiFactSum to evaluate the performance of SlotSum. Experimental results demonstrate that SlotSum could generate summaries that are significantly more factual with credible external knowledge.