arXiv subjects

Robert West

Publications and source records attributed to Robert West.

At least 37 records · Page 2Linked to original sources

arXiv2025

From Model Training to Model Raising

Current AI training methods align models with human values only after their core capabilities have been established, resulting in models that are easily misaligned and lack deep-rooted value systems. We propose a paradigm shift from "model training" to "model raising", in which alignment is woven into a model's development from the start. We identify several key components for this paradigm, all centered around redesigning the training corpus: reframing training data from a first-person perspective, recontextualizing information as lived experience, simulating social interactions, and scaffolding the ordering of training data. We expect that this redesign of the training corpus will lead to an early commitment to values from the first training token onward, such that knowledge, skills, and values are intrinsically much harder to separate. In an ecosystem in which large language model capabilities start overtaking human capabilities in many tasks, this seems to us like a critical need.

cs.AI↗

arXiv2025

One-Step is Enough: Sparse Autoencoders for Text-to-Image Diffusion Models

For large language models (LLMs), sparse autoencoders (SAEs) have been shown to decompose intermediate representations that often are not interpretable directly into sparse sums of interpretable features, facilitating better control and subsequent analysis. However, similar analyses and approaches have been lacking for text-to-image models. We investigate the possibility of using SAEs to learn interpretable features for SDXL Turbo, a few-step text-to-image diffusion model. To this end, we train SAEs on the updates performed by transformer blocks within SDXL Turbo's denoising U-net in its 1-step setting. Interestingly, we find that they generalize to 4-step SDXL Turbo and even to the multi-step SDXL base model (i.e., a different model) without additional training. In addition, we show that their learned features are interpretable, causally influence the generation process, and reveal specialization among the blocks. We do so by creating RIEBench, a representation-based image editing benchmark, for editing images while they are generated by turning on and off individual SAE features. This allows us to track which transformer blocks' features are the most impactful depending on the edit category. Our work is the first investigation of SAEs for interpretability in text-to-image diffusion models and our results establish SAEs as a promising approach for understanding and manipulating the internal mechanisms of text-to-image models.

cs.LG↗

arXiv2025

zip2zip: Inference-Time Adaptive Tokenization via Online Compression

Tokenization efficiency plays a critical role in the performance and cost of large language models (LLMs), yet most models rely on static tokenizers optimized on general-purpose corpora. These tokenizers' fixed vocabularies often fail to adapt to domain- or language-specific inputs, leading to longer token sequences and higher computational costs. We introduce zip2zip, a novel method for achieving context-adaptive tokenization in LLMs at inference time. Leveraging an online data compression algorithm (Lempel-Ziv-Welch), zip2zip dynamically expands its active vocabulary at inference time by continuously replacing fragmented token sequences with more compact hypertokens, which it can immediately output during generation. In doing so, the model refines its internal tokenization scheme to match the token distribution of the current context, reducing redundancy and improving representational efficiency. zip2zip consists of three key components: (1) a tokenizer based on Lempel-Ziv-Welch compression that incrementally merges co-occurring tokens into reusable hypertokens on the fly; (2) a dynamic embedding (and unembedding) layer that computes embeddings for newly formed hypertokens at runtime; and (3) a variant of autoregressive language modeling that pretrains the model to handle hypertokenized, compressed text sequences as inputs and outputs. We show that an existing LLM can be uptrained for zip2zip in 10 GPU-hours via parameter-efficient finetuning. The resulting LLM performs test-time adaptation, learning to use hypertokens in unseen contexts and reducing input and output tokens by 15-40%.

cs.CL↗

arXiv2025

Valid Survey Simulations with Limited Human Data: The Roles of Prompting, Fine-Tuning, and Rectification

Surveys provide valuable insights into public opinion and behavior, but their execution is costly and slow. Large language models (LLMs) have been proposed as a scalable, low-cost substitute for human respondents, but their outputs are often biased and yield invalid estimates. We study the interplay between synthesis methods that use LLMs to generate survey responses and rectification methods that debias population estimates, and explore how human responses are best allocated between them. Using two panel surveys with questions on nutrition, politics, and economics, we find that synthesis alone introduces substantial bias (24-86%), whereas combining it with rectification reduces bias below 5% and increases effective sample size by up to 14%. Overall, we challenge the common practice of using all human responses for fine-tuning, showing that under a fixed budget, allocating most to rectification results in far more effective estimation.

cs.CL↗

arXiv2025

GRAD: Generative Retrieval-Aligned Demonstration Sampler for Efficient Few-Shot Reasoning

Large Language Models (LLMs) achieve strong performance across diverse tasks, but their effectiveness often depends on the quality of the provided context. Retrieval-Augmented Generation (RAG) enriches prompts with external information, but its reliance on static databases constrains adaptability and can result in irrelevant demonstrations. In this work, we propose a Generative Retrieval-Aligned Demonstrator (GRAD), a dynamic demonstration-based approach where an LLM model is trained to generate input-specific concise demonstrations. By tailoring demonstrations to each input, our method offers better contextual support than traditional RAG approaches. We demonstrate the superiority of GRAD under budget constraints, where we limit both the number of tokens used per demonstration and the number of tokens used for the final output. Trained solely on a math dataset, GRAD consistently outperforms strong baselines on Qwen2.5-14B across mathematical reasoning and advanced STEM questions, highlighting GRAD's robust generalization to out-of-distribution (OOD) domains such as physics, chemistry, and computer science. Furthermore, we show that demonstrations generated by trained smaller models can effectively guide larger target models, reducing training costs while maintaining competitive accuracy. Overall, this work introduces a scalable demonstration generator model presenting the first step toward a dynamic few-shot learning paradigm in resource-constrained settings. We release the code used for the project.

cs.CL↗

arXiv2025

Combining Constrained and Unconstrained Decoding via Boosting: BoostCD and Its Application to Information Extraction

Many recent approaches to structured NLP tasks use an autoregressive language model $M$ to map unstructured input text $x$ to output text $y$ representing structured objects (such as tuples, lists, trees, code, etc.), where the desired output structure is enforced via constrained decoding. During training, these approaches do not require the model to be aware of the constraints, which are merely implicit in the training outputs $y$. This is advantageous as it allows for dynamic constraints without requiring retraining, but can lead to low-quality output during constrained decoding at test time. We overcome this problem with Boosted Constrained Decoding (BoostCD), which combines constrained and unconstrained decoding in two phases: Phase 1 decodes from the base model $M$ twice, in constrained and unconstrained mode, obtaining two weak predictions. In phase 2, a learned autoregressive boosted model combines the two weak predictions into one final prediction. The mistakes made by the base model with vs. without constraints tend to be complementary, which the boosted model learns to exploit for improved performance. We demonstrate the power of BoostCD by applying it to closed information extraction. Our model, BoostIE, outperforms prior approaches both in and out of distribution, addressing several common errors identified in those approaches.

cs.CL↗

arXiv2025

Interactive Evaluation of Large Language Models for Multi-Requirement Software Engineering Tasks

Standard single-turn, static benchmarks fall short in evaluating the nuanced capabilities of Large Language Models (LLMs) on complex tasks such as software engineering. In this work, we propose a novel interactive evaluation framework that assesses LLMs on multi-requirement programming tasks through structured, feedback-driven dialogue. Each task is modeled as a requirement dependency graph, and an ``interviewer'' LLM, aware of the ground-truth solution, provides minimal, targeted hints to an ``interviewee'' model to help correct errors and fulfill target constraints. This dynamic protocol enables fine-grained diagnostic insights into model behavior, uncovering strengths and systematic weaknesses that static benchmarks fail to measure. We build on DevAI, a benchmark of 55 curated programming tasks, by adding ground-truth solutions and evaluating the relevance and utility of interviewer hints through expert annotation. Our results highlight the importance of dynamic evaluation in advancing the development of collaborative code-generating agents.

cs.AI↗

arXiv2025

Assessing Social Alignment: Do Personality-Prompted Large Language Models Behave Like Humans?

The ongoing revolution in language modeling has led to various novel applications, some of which rely on the emerging social abilities of large language models (LLMs). Already, many turn to the new cyber friends for advice during the pivotal moments of their lives and trust them with the deepest secrets, implying that accurate shaping of the LLM's personality is paramount. To this end, state-of-the-art approaches exploit a vast variety of training data, and prompt the model to adopt a particular personality. We ask (i) if personality-prompted models behave (i.e., make decisions when presented with a social situation) in line with the ascribed personality (ii) if their behavior can be finely controlled. We use classic psychological experiments, the Milgram experiment and the Ultimatum Game, as social interaction testbeds and apply personality prompting to open- and closed-source LLMs from 4 different vendors. Our experiments reveal failure modes of the prompt-based modulation of the models' behavior that are shared across all models tested and persist under prompt perturbations. These findings challenge the optimistic sentiment toward personality prompting generally held in the community.

cs.CY↗

arXiv2025

Quotegraph: A Social Network Extracted from Millions of News Quotations

We introduce Quotegraph, a novel large-scale social network derived from speaker-attributed quotations in English news articles published between 2008 and 2020. Quotegraph consists of 528 thousand unique nodes and 8.63 million directed edges, pointing from speakers to persons they mention. The nodes are linked to their corresponding items in Wikidata, thereby endowing the dataset with detailed biographic entity information, including nationality, gender, and political affiliation. Being derived from Quotebank, a massive corpus of quotations, relations in Quotegraph are additionally enriched with the information about the context in which they are featured. Each part of the network construction pipeline is language agnostic, enabling the construction of similar datasets based on non-English news corpora. We believe Quotegraph is a compelling resource for computational social scientists, complementary to online social networks, with the potential to yield novel insights into the behavior of public figures and how it is captured in the news.

cs.SI↗

arXiv2025

TRPrompt: Bootstrapping Query-Aware Prompt Optimization from Textual Rewards

Prompt optimization improves the reasoning abilities of large language models (LLMs) without requiring parameter updates to the target model. Following heuristic-based "Think step by step" approaches, the field has evolved in two main directions: while one group of methods uses textual feedback to elicit improved prompts from general-purpose LLMs in a training-free way, a concurrent line of research relies on numerical rewards to train a special prompt model, tailored for providing optimal prompts to the target model. In this paper, we introduce the Textual Reward Prompt framework (TRPrompt), which unifies these approaches by directly incorporating textual feedback into training of the prompt model. Our framework does not require prior dataset collection and is being iteratively improved with the feedback on the generated prompts. When coupled with the capacity of an LLM to internalize the notion of what a "good" prompt is, the high-resolution signal provided by the textual rewards allows us to train a prompt model yielding state-of-the-art query-specific prompts for the problems from the challenging math datasets GSMHard and MATH.

cs.CL↗

arXiv2025

Separating Tongue from Thought: Activation Patching Reveals Language-Agnostic Concept Representations in Transformers

A central question in multilingual language modeling is whether large language models (LLMs) develop a universal concept representation, disentangled from specific languages. In this paper, we address this question by analyzing latent representations (latents) during a word-translation task in transformer-based LLMs. We strategically extract latents from a source translation prompt and insert them into the forward pass on a target translation prompt. By doing so, we find that the output language is encoded in the latent at an earlier layer than the concept to be translated. Building on this insight, we conduct two key experiments. First, we demonstrate that we can change the concept without changing the language and vice versa through activation patching alone. Second, we show that patching with the mean representation of a concept across different languages does not affect the models' ability to translate it, but instead improves it. Finally, we generalize to multi-token generation and demonstrate that the model can generate natural language description of those mean representations. Our results provide evidence for the existence of language-agnostic concept representations within the investigated models.

cs.CL↗

arXiv2025

Controllable Context Sensitivity and the Knob Behind It

When making predictions, a language model must trade off how much it relies on its context vs. its prior knowledge. Choosing how sensitive the model is to its context is a fundamental functionality, as it enables the model to excel at tasks like retrieval-augmented generation and question-answering. In this paper, we search for a knob which controls this sensitivity, determining whether language models answer from the context or their prior knowledge. To guide this search, we design a task for controllable context sensitivity. In this task, we first feed the model a context (Paris is in England) and a question (Where is Paris?); we then instruct the model to either use its prior or contextual knowledge and evaluate whether it generates the correct answer for both intents (either France or England). When fine-tuned on this task, instruction-tuned versions of Llama-3.1, Mistral-v0.3, and Gemma-2 can solve it with high accuracy (85-95%). Analyzing these high-performing models, we narrow down which layers may be important to context sensitivity using a novel linear time algorithm. Then, in each model, we identify a 1-D subspace in a single layer that encodes whether the model follows context or prior knowledge. Interestingly, while we identify this subspace in a fine-tuned model, we find that the exact same subspace serves as an effective knob in not only that model but also non-fine-tuned instruct and base models of that model family. Finally, we show a strong correlation between a model's performance and how distinctly it separates context-agreeing from context-ignoring answers in this subspace. These results suggest a single subspace facilitates how the model chooses between context and prior knowledge, hinting at a simple fundamental mechanism that controls this behavior.

cs.CL↗

arXiv2025

Web2Wiki: Characterizing Wikipedia Linking Across the Web

Wikipedia is one of the most visited websites globally, yet its role beyond its own platform remains largely unexplored. In this paper, we present the first large-scale analysis of how Wikipedia is referenced across the Web. Using a dataset from Common Crawl, we identify over 90 million Wikipedia links spanning 1.68% of Web domains and examine their distribution, context, and function. Our analysis of English Wikipedia reveals three key findings: (1) Wikipedia is most frequently cited by news and science websites for informational purposes, while commercial websites reference it less often. (2) The majority of Wikipedia links appear within the main content rather than in boilerplate or user-generated sections, highlighting their role in structured knowledge presentation. (3) Most links (95%) serve as explanatory references rather than as evidence or attribution, reinforcing Wikipedia's function as a background knowledge provider. While this study focuses on English Wikipedia, our publicly released Web2Wiki dataset includes links from multiple language editions, supporting future research on Wikipedia's global influence on the Web.

cs.SI↗

arXiv2025

Fleet of Agents: Coordinated Problem Solving with Large Language Models

While numerous frameworks have been developed to enhance the reasoning abilities of large language models (LLMs), there is a scarcity of methods that effectively balance the trade-off between cost and quality. In this paper, we introduce Fleet of Agents (FoA), a novel and intuitive yet principled framework utilizing LLMs as agents to navigate through dynamic tree searches, employing a genetic-type particle filtering approach. FoA spawns a multitude of agents, each exploring the search space autonomously, followed by a selection phase where resampling based on a heuristic value function optimizes the balance between exploration and exploitation. This mechanism enables dynamic branching, adapting the exploration strategy based on discovered solutions. We conduct extensive experiments on three benchmark tasks, ``Game of 24'', ``Mini-Crosswords'', and ``WebShop'', utilizing four different LLMs, ``GPT-3.5'', ``GPT-4'', ``LLaMA3.2-11B'', and ``LLaMA3.2-90B''. On average across all tasks and LLMs, FoA obtains a quality improvement of ~5% while requiring only ~40% of the cost of previous SOTA methods. Notably, our analyses reveal that (1) FoA achieves the best cost-quality trade-off among all benchmarked methods and (2) FoA + LLaMA3.2-11B surpasses the Llama3.2-90B model. FoA is publicly available at https://github.com/au-clan/FoA.

cs.CL↗

arXiv2025

A Logical Fallacy-Informed Framework for Argument Generation

Despite the remarkable performance of Large Language Models (LLMs) in natural language processing tasks, they still struggle with generating logically sound arguments, resulting in potential risks such as spreading misinformation. To address this issue, we introduce FIPO, a fallacy-informed framework that leverages preference optimization methods to steer LLMs toward logically sound arguments. FIPO includes a classification loss, to capture the fine-grained information on fallacy types. Our results on argumentation datasets show that our method reduces the fallacy errors by up to 17.5%. Furthermore, our human evaluation results indicate that the quality of the generated arguments by our method significantly outperforms the fine-tuned baselines, as well as other preference optimization methods, such as DPO. These findings highlight the importance of ensuring models are aware of logical fallacies for effective argument generation. Our code is available at github.com/lucamouchel/Logical-Fallacies.

cs.CL↗

arXiv2025

Agentic AI: The Era of Semantic Decoding

Recent work demonstrated great promise in the idea of orchestrating collaborations between LLMs, human input, and various tools to address the inherent limitations of LLMs. We propose a novel perspective called semantic decoding, which frames these collaborative processes as optimization procedures in semantic space. Specifically, we conceptualize LLMs as semantic processors that manipulate meaningful pieces of information that we call semantic tokens (known thoughts). LLMs are among a large pool of other semantic processors, including humans and tools, such as search engines or code executors. Collectively, semantic processors engage in dynamic exchanges of semantic tokens to progressively construct high-utility outputs. We refer to these orchestrated interactions among semantic processors, optimizing and searching in semantic space, as semantic decoding algorithms. This concept draws a direct parallel to the well-studied problem of syntactic decoding, which involves crafting algorithms to best exploit auto-regressive language models for extracting high-utility sequences of syntactic tokens. By focusing on the semantic level and disregarding syntactic details, we gain a fresh perspective on the engineering of AI systems, enabling us to imagine systems with much greater complexity and capabilities. In this position paper, we formalize the transition from syntactic to semantic tokens as well as the analogy between syntactic and semantic decoding. Subsequently, we explore the possibilities of optimizing within the space of semantic tokens via semantic decoding algorithms. We conclude with a list of research opportunities and questions arising from this fresh perspective. The semantic decoding perspective offers a powerful abstraction for search and optimization directly in the space of meaningful concepts, with semantic tokens as the fundamental units of a new type of computation.

cs.CL↗

arXiv2025

Does Content Moderation Lead Users Away from Fringe Movements? Evidence from a Recovery Community

Online platforms have sanctioned individuals and communities associated with fringe movements linked to hate speech, violence, and terrorism, but can these sanctions contribute to the abandonment of these movements? Here, we investigate this question through the lens of exredpill, a recovery community on Reddit meant to help individuals leave movements within the Manosphere, a conglomerate of fringe Web based movements focused on men's issues. We conduct an observational study on the impact of sanctioning some of Reddit's largest Manosphere communities on the activity levels and user influx of exredpill, the largest associated recovery subreddit. We find that banning a related radical community positively affects participation in exredpill in the period following the ban. Yet, quarantining the community, a softer moderation intervention, yields no such effects. We show that the effect induced by banning a radical community is stronger than for some of the widely discussed real-world events related to the Manosphere and that moderation actions against the Manosphere do not cause a spike in toxicity or malicious activity in exredpill. Overall, our findings suggest that content moderation acts as a deradicalization catalyst.

cs.SI↗

arXiv2025

Localized Cultural Knowledge is Conserved and Controllable in Large Language Models

Just as humans display language patterns influenced by their native tongue when speaking new languages, LLMs often default to English-centric responses even when generating in other languages. Nevertheless, we observe that local cultural information persists within the models and can be readily activated for cultural customization. We first demonstrate that explicitly providing cultural context in prompts significantly improves the models' ability to generate culturally localized responses. We term the disparity in model performance with versus without explicit cultural context the explicit-implicit localization gap, indicating that while cultural knowledge exists within LLMs, it may not naturally surface in multilingual interactions if cultural context is not explicitly provided. Despite the explicit prompting benefit, however, the answers reduce in diversity and tend toward stereotypes. Second, we identify an explicit cultural customization vector, conserved across all non-English languages we explore, which enables LLMs to be steered from the synthetic English cultural world-model toward each non-English cultural world. Steered responses retain the diversity of implicit prompting and reduce stereotypes to dramatically improve the potential for customization. We discuss the implications of explicit cultural customization for understanding the conservation of alternative cultural world models within LLMs, and their controllable utility for translation, cultural customization, and the possibility of making the explicit implicit through soft control for expanded LLM function and appeal.

cs.CL↗