Large language models (LLMs) are increasingly integral to information retrieval (IR), powering ranking, evaluation, and AI-assisted content creation. This widespread adoption necessitates a critical examination of potential biases arising from the interplay between these LLM-based components. This paper synthesizes existing research and presents novel experiment designs that explore how LLM-based rankers and assistants influence LLM-based judges. We provide the first empirical evidence of LLM judges exhibiting significant bias towards LLM-based rankers. Furthermore, we observe limitations in LLM judges' ability to discern subtle system performance differences. Contrary to some previous findings, our preliminary study does not find evidence of bias against AI-generated content. These results highlight the need for a more holistic view of the LLM-driven information ecosystem. To this end, we offer initial guidelines and a research agenda to ensure the reliable use of LLMs in IR evaluation.

Human understanding of language is robust to different word choices as far as they represent similar semantic concepts. To what extent does our human intuition transfer to language models, which represent all subwords as distinct embeddings? In this work, we take an initial step on measuring the role of shared semantics among subwords in the encoder-only multilingual language models (mLMs). To this end, we form "semantic tokens" by merging the semantically similar subwords and their embeddings, and evaluate the updated mLMs on 5 heterogeneous multilingual downstream tasks. Results show that the general shared semantics could get the models a long way in making the predictions on mLMs with different tokenizers and model sizes. Inspections on the grouped subwords show that they exhibit a wide range of semantic similarities, including synonyms and translations across many languages and scripts. Lastly, we found the zero-shot results with semantic tokens are on par or even better than the original models on certain classification tasks, suggesting that the shared subword-level semantics may serve as the anchors for cross-lingual transferring.

Watermarking has recently emerged as an effective strategy for detecting the outputs of large language models (LLMs). Most existing schemes require whitebox access to the model's next-token probability distribution, which is typically not accessible to downstream users of an LLM API. In this work, we propose a principled watermarking scheme that requires only the ability to sample sequences from the LLM (i.e. We provide performance guarantees, demonstrate how it can be leveraged when white-box access is available, and show when it can outperform existing white-box schemes via comprehensive experiments. It can be critical to understand whether a piece of text is generated by a large language model (LLM). For instance, one often wants to know how trustworthy a piece of text is, and those written by an LLM may be deemed untrustworthy as these models can hallucinate. This problem comes in different flavors -- one may want to detect whether it was generated by a specific model or by any model. Furthermore, the detecting party may or may not have white-box access (e.g. an ability to compute log-probabilities) to the generator they wish to test against. Typically, parties that have white-box access are the owners of the model so we refer to this case as first-party detection and the counterpart as third-party detection. The goal of watermarking is to cleverly bias the generator so that first-party detection becomes easier. Most proposed techniques do not modify the underlying LLM's model weights or its training procedure but rather inject the watermark during autoregressive decoding at inference time. They require access to the next-token logits and inject the watermark every step of the sampling loop. This required access prevents third-party users of an LLM from applying their own watermark as proprietary APIs currently do not support this option. Supporting this functionality presents a security risk in addition to significant engineering considerations.

A key requirement in developing Generative Language Models (GLMs) is to have their values aligned with human values. Preference-based alignment is a widely used paradigm for this purpose, in which preferences over generation pairs are first elicited from human annotators or AI systems, and then fed into some alignment techniques, e.g., Direct Preference Optimization. However, a substantial percent (20 - 40%) of the preference pairs used in GLM alignment are noisy, and it remains unclear how the noise affects the alignment performance and how to mitigate its negative impact. In this paper, we propose a framework to inject desirable amounts and types of noise to the preferences, and systematically study the impact of preference noise on the alignment performance in two tasks (summarization and dialogue generation). We find that the alignment performance can be highly sensitive to the noise rates in the preference data: e.g., a 10 percentage points (pp) increase of the noise rate can lead to 30 pp drop in the alignment performance (in win rate). To mitigate the impact of noise, confidence-based data filtering shows significant benefit when certain types of noise are present. We hope our work can help the community better understand and mitigate the impact of preference noise in GLM alignment.

Reinforcement Learning from Human Feedback (RLHF) is a popular method for aligning Language Models (LM) with human values and preferences. RLHF requires a large number of preference pairs as training data, which are often used in both the Supervised Fine-Tuning and Reward Model training, and therefore publicly available datasets are commonly used. In this work, we study to what extent a malicious actor can manipulate the LMs generations by poisoning the preferences, i.e., injecting poisonous preference pairs into these datasets and the RLHF training process. We propose strategies to build poisonous preference pairs and test their performance by poisoning two widely used preference datasets. Our results show that preference poisoning is highly effective: by injecting a small amount of poisonous data (1-5% of the original dataset), we can effectively manipulate the LM to generate a target entity in a target sentiment (positive or negative). The findings from our experiments also shed light on strategies to defend against the preference poisoning attack.

Differentiable Search Indices (DSIs) encode a corpus of documents in model parameters and use the same model to answer user queries directly. Despite the strong performance of DSI models, deploying them in situations where the corpus changes over time is computationally expensive because reindexing the corpus requires re-training the model. In this work, we introduce DSI++, a continual learning challenge for DSI to incrementally index new documents while being able to answer queries related to both previously and newly indexed documents. Across different model scales and document identifier representations, we show that continual indexing of new documents leads to considerable forgetting of previously indexed documents. We also hypothesize and verify that the model experiences forgetting events during training, leading to unstable learning. To mitigate these issues, we investigate two approaches. The first focuses on modifying the training dynamics. Flatter minima implicitly alleviate forgetting, so we optimize for flatter loss basins and show that the model stably memorizes more documents ($+12\%$). Next, we introduce a generative memory to sample pseudo-queries for documents and supplement them during continual indexing to prevent forgetting for the retrieval task. Extensive experiments on novel continual indexing benchmarks based on Natural Questions (NQ) and MS MARCO demonstrate that our proposed solution mitigates forgetting significantly. Concretely, it improves the average Hits@10 by $+21.1\%$ over competitive baselines for NQ and requires $6$ times fewer model updates compared to re-training the DSI model for incrementally indexing five corpora in a sequence.

Recently proposed long-form question answering (QA) systems, supported by large language models (LLMs), have shown promising capabilities. Yet, attributing and verifying their generated abstractive answers can be difficult, and automatically evaluating their accuracy remains an ongoing challenge. In this work, we introduce a new QA task for answering multi-answer questions by summarizing multiple diverse sources in a semi-extractive fashion. Specifically, Semi-extractive Multi-source QA (SEMQA) requires models to output a comprehensive answer, while mixing factual quoted spans -- copied verbatim from given input sources -- and non-factual free-text connectors that glue these spans together into a single cohesive passage. This setting bridges the gap between the outputs of well-grounded but constrained extractive QA systems and more fluent but harder to attribute fully abstractive answers. Particularly, it enables a new mode for language models that leverages their advanced language generation capabilities, while also producing fine in-line attributions by-design that are easy to verify, interpret, and evaluate. To study this task, we create the first dataset of this kind, QuoteSum, with human-written semi-extractive answers to natural and generated questions, and define text-based evaluation metrics. Experimenting with several LLMs in various settings, we find this task to be surprisingly challenging, demonstrating the importance of QuoteSum for developing and studying such consolidation capabilities.

Ranking documents using Large Language Models (LLMs) by directly feeding the query and candidate documents into the prompt is an interesting and practical problem. However, there has been limited success so far, as researchers have found it difficult to outperform fine-tuned baseline rankers on benchmark datasets. We analyze pointwise and listwise ranking prompts used by existing methods and argue that off-the-shelf LLMs do not fully understand these ranking formulations, possibly due to the nature of how LLMs are trained. In this paper, we propose to significantly reduce the burden on LLMs by using a new technique called Pairwise Ranking Prompting (PRP). Our results are the first in the literature to achieve state-of-the-art ranking performance on standard benchmarks using moderate-sized open-sourced LLMs. On TREC-DL2020, PRP based on the Flan-UL2 model with 20B parameters outperforms the previous best approach in the literature, which is based on the blackbox commercial GPT-4 that has 50x (estimated) model size, by over 5% at NDCG@1. On TREC-DL2019, PRP is only inferior to the GPT-4 solution on the NDCG@5 and NDCG@10 metrics, while outperforming other existing solutions, such as InstructGPT which has 175B parameters, by over 10% for nearly all ranking metrics. Furthermore, we propose several variants of PRP to improve efficiency and show that it is possible to achieve competitive results even with linear complexity. We also discuss other benefits of PRP, such as supporting both generation and scoring LLM APIs, as well as being insensitive to input ordering.

Generative information retrieval (IR) has experienced substantial growth across multiple research communities (e.g., information retrieval, computer vision, natural language processing, and machine learning), and has been highly visible in the popular press. Theoretical, empirical, and actual user-facing products have been released that retrieve documents (via generation) or directly generate answers given an input request. We would like to investigate whether end-to-end generative models are just another trend or, as some claim, a paradigm change for IR. This necessitates new metrics, theoretical grounding, evaluation methods, task definitions, models, user interfaces, etc. The goal of this workshop (https://coda.io/@sigir/gen-ir) is to focus on previously explored Generative IR techniques like document retrieval and direct Grounded Answer Generation, while also offering a venue for the discussion and exploration of how Generative IR can be applied to new domains like recommendation systems, summarization, etc. The format of the workshop is interactive, including roundtable and keynote sessions and tends to avoid the one-sided dialogue of a mini-conference.

Transformer encoders contextualize token representations by attending to all other tokens at each layer, leading to quadratic increase in compute effort with the input length. In practice, however, the input text of many NLP tasks can be seen as a sequence of related segments (e.g., the sequence of sentences within a passage, or the hypothesis and premise in NLI). While attending across these segments is highly beneficial for many tasks, we hypothesize that this interaction can be delayed until later encoding stages. To this end, we introduce Layer-Adjustable Interactions in Transformers (LAIT). Within LAIT, segmented inputs are first encoded independently, and then jointly. This partial two-tower architecture bridges the gap between a Dual Encoder's ability to pre-compute representations for segments and a fully self-attentive Transformer's capacity to model cross-segment attention. The LAIT framework effectively leverages existing pretrained Transformers and converts them into the hybrid of the two aforementioned architectures, allowing for easy and intuitive control over the performance-efficiency tradeoff. Experimenting on a wide range of NLP tasks, we find LAIT able to reduce 30-50% of the attention FLOPs on many tasks, while preserving high accuracy; in some practical settings, LAIT could reduce actual latency by orders of magnitude.