We find that the cross-entropy loss curves of neural language models empirically adhere to a scaling law with learning rate (LR) annealing over training steps ($s$): $$L(s) = L_0 + A\cdot S_1^{-\alpha} - C\cdot S_2$$ Where $S_1$ is forward area and $S_2$ is learning rate annealing area. This formulation takes into account two factors: (1) The forward scaling defined as typical scaling law, and (2) the additional loss drop brought by LR annealing. Therefore, this formulation can describe the full loss curve at each step, rather than the single loss point at the end of training. Applying the scaling law with LR annealing and fitting only one or two training curves, we can accurately predict the loss of language model training at any given step and across any learning rate scheduler (LRS). Furthermore, this equation accurately describes the dynamics during training process, and provides a theoretical verification and explanation for numerous experimental findings of previous studies, particularly those focusing on LR schedule and LR annealing. The resulting insights, also serve as a guide for researchers to select critical LRS in advance by prediction using our equation. Most significantly, since all the points in a full training curve follow the equation, we can achieve accurate loss prediction at any given step across any learning rate scheduler, while expending less than 1\% of the computational cost required by the chinchilla scaling law to fit language modeling loss. This approach extremely democratizes scaling law fitting and predicting in developing large language models.

Abstract--Large Language Models (LLMs) have seen increasing use in various software development tasks, especially in code generation. The most advanced recent methods attempt to incorporate feedback from code execution into prompts to help guide LLMs in generating correct code, in an iterative process. While effective, these methods could be costly and time-consuming due to numerous interactions with the LLM and the extensive token usage. To address this issue, we propose an alternative approach named Evolutionary Prompt Engineering for Code (EPiC), which leverages a lightweight evolutionary algorithm to evolve the original prompts toward better ones that produce high-quality code, with minimal interactions with LLM. Our evaluation against state-of-the-art (SOTA) LLM-based code generation models shows that EPiC outperforms all the baselines in terms of cost-effectiveness. Among these activities, code generation using LLMs has demonstrated significant potential. In LLM-based code generation, various prompt engineering techniques, including zero-shot [5], in-context learning [33], [34], RAG [35], and task-specific methods [36], [37], have been shown to outperform fine-tuned smaller models. The most advanced prompt engineering methods for code generation employ various agent-based approaches [28]. SOTA methods such as Reflexion [20], Language Agent Tree Search (LATS) [21], AgentCoder [22], LDB Figure 1: The initial failed prompt (left) and the mutated [23], and MetaGPT [29] are either planning-based or multicollaborative successful prompt (right) agents. While effective, these methods can be costly and time-consuming due to numerous interactions with LLMs which results in extensive token usage, making them less attractive in practical settings. The first phase involves primary code requires on average 3 minutes to generate the implementation generation using an initial prompt and its evaluation using of a function with an average of only 6 lines of code, on a set of test cases. If a correct solution is not generated, the MBPP dataset.

The utility of large language models (LLMs) depends heavily on the quality and quantity of their training data. Many organizations possess large data corpora that could be leveraged to train or fine-tune LLMs tailored to their specific needs. However, these datasets often come with access restrictions that are based on user privileges and enforced by access control mechanisms. Training LLMs on such datasets could result in exposure of sensitive information to unauthorized users. A straightforward approach for preventing such exposure is to train a separate model for each access level. This, however, may result in low utility models due to the limited amount of training data per model compared to the amount in the entire organizational corpus. Another approach is to train a single LLM on all the data while limiting the exposure of unauthorized information. However, current exposure-limiting methods for LLMs are ineffective for access-controlled data, where sensitive information appears frequently across many training examples. We propose DOMBA - double model balancing - a simple approach for training and deploying LLMs that provides high utility and access-control functionality with security guarantees. DOMBA aggregates the probability distributions of two models, each trained on documents with (potentially many) different access levels, using a "min-bounded" average function (a function that is bounded by the smaller value, e.g., harmonic mean). A detailed mathematical analysis and extensive evaluation show that DOMBA safeguards restricted information while offering utility comparable to non-secure models.

We introduce a group of related methods for binary classification tasks using probes of the hidden state activations in large language models (LLMs). Performance is on par with the largest and most advanced LLMs currently available, but requiring orders of magnitude fewer computational resources and not requiring labeled data. This approach involves translating class labels into a semantically rich description, spontaneous symmetry breaking of multilayer perceptron probes for unsupervised learning and inference, training probes to generate confidence scores (prior probabilities) from hidden state activations subject to known constraints via entropy maximization, and selecting the most confident probe model from an ensemble for prediction. These techniques are evaluated on four datasets using five base LLMs.

Despite the widespread adoption of autoregressive language models, explainability evaluation research has predominantly focused on span infilling and masked language models (MLMs). Evaluating the faithfulness of an explanation method -- how accurately the method explains the inner workings and decision-making of the model -- is very challenging because it is very hard to separate the model from its explanation. Most faithfulness evaluation techniques corrupt or remove some input tokens considered important according to a particular attribution (feature importance) method and observe the change in the model's output. This approach creates out-of-distribution inputs for causal language models (CLMs) due to their training objective of next token prediction. In this study, we propose a technique that leverages counterfactual generation to evaluate the faithfulness of attribution methods for autoregressive language modeling scenarios. Our technique creates fluent and in-distribution counterfactuals that makes evaluation protocol more reliable. Code is available at https://github.com/Sepehr-Kamahi/faith

The fashion domain encompasses a variety of real-world multimodal tasks, including multimodal retrieval and multimodal generation. The rapid advancements in artificial intelligence generated content, particularly in technologies like large language models for text generation and diffusion models for visual generation, have sparked widespread research interest in applying these multimodal models in the fashion domain. However, tasks involving embeddings, such as image-to-text or text-to-image retrieval, have been largely overlooked from this perspective due to the diverse nature of the multimodal fashion domain. And current research on multi-task single models lack focus on image generation. In this work, we present UniFashion, a unified framework that simultaneously tackles the challenges of multimodal generation and retrieval tasks within the fashion domain, integrating image generation with retrieval tasks and text generation tasks. UniFashion unifies embedding and generative tasks by integrating a diffusion model and LLM, enabling controllable and high-fidelity generation. Our model significantly outperforms previous single-task state-of-the-art models across diverse fashion tasks, and can be readily adapted to manage complex vision-language tasks. This work demonstrates the potential learning synergy between multimodal generation and retrieval, offering a promising direction for future research in the fashion domain. The source code is available at https://github.com/xiangyu-mm/UniFashion.

Despite prior safety alignment efforts, mainstream LLMs can still generate harmful and unethical content when subjected to jailbreaking attacks. Existing jailbreaking methods fall into two main categories: template-based and optimization-based methods. The former requires significant manual effort and domain knowledge, while the latter, exemplified by Greedy Coordinate Gradient (GCG), which seeks to maximize the likelihood of harmful LLM outputs through token-level optimization, also encounters several limitations: requiring white-box access, necessitating pre-constructed affirmative phrase, and suffering from low efficiency. In this paper, we present ECLIPSE, a novel and efficient black-box jailbreaking method utilizing optimizable suffixes. Drawing inspiration from LLMs' powerful generation and optimization capabilities, we employ task prompts to translate jailbreaking goals into natural language instructions. This guides the LLM to generate adversarial suffixes for malicious queries. In particular, a harmfulness scorer provides continuous feedback, enabling LLM self-reflection and iterative optimization to autonomously and efficiently produce effective suffixes. Experimental results demonstrate that ECLIPSE achieves an average attack success rate (ASR) of 0.92 across three open-source LLMs and GPT-3.5-Turbo, significantly surpassing GCG in 2.4 times. Moreover, ECLIPSE is on par with template-based methods in ASR while offering superior attack efficiency, reducing the average attack overhead by 83%.

Formal theorem proving, a field at the intersection of mathematics and computer science, has seen renewed interest with advancements in large language models (LLMs). This paper introduces SubgoalXL, a novel approach that synergizes subgoal-based proofs with expert learning to enhance LLMs' capabilities in formal theorem proving within the Isabelle environment. SubgoalXL addresses two critical challenges: the scarcity of specialized mathematics and theorem-proving data, and the need for improved multi-step reasoning abilities in LLMs. Leveraging the Isabelle environment's advantages in subgoal-based proofs, SubgoalXL achieves a new state-of-the-art performance of 56.1% in Isabelle on the standard miniF2F dataset, marking an absolute improvement of 4.9%. Notably, SubgoalXL successfully solves 41 AMC12, 9 AIME, and 3 IMO problems from miniF2F. These results underscore the effectiveness of maximizing limited data utility and employing targeted guidance for complex reasoning in formal theorem proving, contributing to the ongoing advancement of AI reasoning capabilities. Formal theorem proving, a field at the intersection of mathematics and computer science, has flourished alongside the development of languages like Lean (de Moura et al., 2015) and Isabelle (Paulson, 1994). These two prominent communities have been instrumental in advancing the field's core challenge: mechanizing mathematical reasoning and proof verification (Li et al., 2020).

To address the challenge of automating knowledge discovery from a vast volume of literature, in this paper, we introduce a novel framework based on large language models (LLMs) that combines a progressive ontology prompting (POP) algorithm with a dual-agent system, named LLM-Duo, designed to enhance the automation of knowledge extraction from scientific articles. The POP algorithm utilizes a prioritized breadth-first search (BFS) across a predefined ontology to generate structured prompt templates and action orders, thereby guiding LLMs to discover knowledge in an automatic manner. Additionally, our LLM-Duo employs two specialized LLM agents: an explorer and an evaluator. These two agents work collaboratively and adversarially to enhance the reliability of the discovery and annotation processes. Experiments demonstrate that our method outperforms advanced baselines, enabling more accurate and complete annotations. To validate the effectiveness of our method in real-world scenarios, we employ our method in a case study of speech-language intervention discovery. Our method identifies 2,421 interventions from 64,177 research articles in the speech-language therapy domain. We curate these findings into a publicly accessible intervention knowledge base that holds significant potential to benefit the speech-language therapy community.

While previous chapters focused on recommendation systems (RSs) based on standardized, non-verbal user feedback such as purchases, views, and clicks -- the advent of LLMs has unlocked the use of natural language (NL) interactions for recommendation. This chapter discusses how LLMs' abilities for general NL reasoning present novel opportunities to build highly personalized RSs -- which can effectively connect nuanced and diverse user preferences to items, potentially via interactive dialogues. To begin this discussion, we first present a taxonomy of the key data sources for language-driven recommendation, covering item descriptions, user-system interactions, and user profiles. We then proceed to fundamental techniques for LLM recommendation, reviewing the use of encoder-only and autoregressive LLM recommendation in both tuned and untuned settings. Afterwards, we move to multi-module recommendation architectures in which LLMs interact with components such as retrievers and RSs in multi-stage pipelines. This brings us to architectures for conversational recommender systems (CRSs), in which LLMs facilitate multi-turn dialogues where each turn presents an opportunity not only to make recommendations, but also to engage with the user in interactive preference elicitation, critiquing, and question-answering.