NeuroGenesysThe research we build on.
1,514 papers by other researchers that our work cites, from continual learning and verification to language models, agents and hardware. Each entry links to the paper on arXiv and to where we use it. 301 are cited on published pages of this site; the rest in the research notes behind our studies. Our own papers are in the archive.
- 2024-09arXiv 2409.18025An Adversarial Perspective on Machine Unlearning for AI Safety
Jakub Łucki, Boyi Wei, Yangsibo Huang et al. · Machine learning
Large language models are finetuned to refuse questions about hazardous knowledge, but these protections can often be bypassed. Unlearning methods aim at completely removing hazardous capabilities from models and make them inaccessible to adversaries.
- 2024-09arXiv 2409.16636Training Language Models to Win Debates with Self-Play Improves Judge Accuracy
Samuel Arnesen, David Rein, Julian Michael · Language models & NLP
We test the robustness of debate as a method of scalable oversight by training models to debate with data generated via self-play. In a long-context reading comprehension task, we find that language model based evaluators answer questions more accurately when judging models optimized to win debates.
- 2024-09arXiv 2409.16461Strategies for Improving NL-to-FOL Translation with LLMs: Data Generation, Incremental Fine-Tuning, and Verification
Ramya Keerthy Thatikonda, Jiuzhou Han, Wray Buntine et al. · Language models & NLP
Logical reasoning is a fundamental task in natural language processing that presents significant challenges to Large Language Models (LLMs). The inherent characteristics of logical reasoning makes it well-suited for symbolic representations such as first-order logic (FOL).
- 2024-09arXiv 2409.12183To CoT or not to CoT? Chain-of-thought helps mainly on math and symbolic reasoning
Zayne Sprague, Fangcong Yin, Juan Diego Rodriguez et al. · Language models & NLP
Chain-of-thought (CoT) via prompting is the de facto method for eliciting reasoning capabilities from large language models (LLMs). But for what kinds of tasks is this extra ``thinking'' really helpful?
Cited inWhat Language Buys - 2024-09arXiv 2409.07429Agent Workflow Memory
Zora Zhiruo Wang, Jiayuan Mao, Daniel Fried et al. · Language models & NLP
Despite the potential of language model-based agents to solve real-world tasks such as web navigation, current methods still struggle with long-horizon tasks with complex action trajectories.
Cited inInternal notes - 2024-09arXiv 2409.07431Synthetic continued pretraining
Zitong Yang, Neil Band, Shuangping Li et al. · Machine learning
Pretraining on large-scale, unstructured internet text enables language models to acquire a significant amount of world knowledge. However, this knowledge acquisition is data-inefficient--to learn a given fact, models must be trained on hundreds to thousands of diverse representations of it.
Cited inCoherence is selected, anchored, then pruned: a literature synthesis on logical coherence in LLM world modelsStudy: Continual learning order and companion dataStudy: Generational competitive continual learningStudy: Logical coherence of LLM world modelsStudy: Orchestrated specialists and bridge synthesis - 2024-09arXiv 2409.13724Logically Consistent Language Models via Neuro-Symbolic Integration
Diego Calanzone, Stefano Teso, Antonio Vergari · Language models & NLP
Large language models (LLMs) are a promising venue for natural language understanding and generation.
- 2024-09arXiv 2409.03992Confidential Computing on NVIDIA Hopper GPUs: A Performance Benchmark Study
Jianwei Zhu, Hang Yin, Peng Deng et al. · Other
This report evaluates the performance impact of enabling Trusted Execution Environments (TEE) on NVIDIA Hopper GPUs for large language model (LLM) inference tasks.
- 2024-09arXiv 2409.02060OLMoE: Open Mixture-of-Experts Language Models
Niklas Muennighoff, Luca Soldaini, Dirk Groeneveld et al. · Language models & NLP
We introduce OLMoE, a fully open, state-of-the-art language model leveraging sparse Mixture-of-Experts (MoE). OLMoE-1B-7B has 7 billion (B) parameters but uses only 1B per input token.
- 2024-08arXiv 2408.15664Auxiliary-Loss-Free Load Balancing Strategy for Mixture-of-Experts
Lean Wang, Huazuo Gao, Chenggang Zhao et al. · Machine learning
For Mixture-of-Experts (MoE) models, an unbalanced expert load will lead to routing collapse or increased computational overhead.
- 2024-08arXiv 2408.15901Nexus: Specialization meets Adaptability for Efficiently Training Mixture of Experts
Nikolas Gritsch, Qizhen Zhang, Acyr Locatelli et al. · Language models & NLP
Efficiency, specialization, and adaptability to new data distributions are qualities that are hard to combine in current Large Language Models.
- 2024-08arXiv 2408.14677Can Optimization Trajectories Explain Multi-Task Transfer?
David Mueller, Mark Dredze, Nicholas Andrews · Machine learning
Despite the widespread adoption of multi-task training in deep learning, little is understood about how multi-task learning (MTL) affects generalization.
- 2024-08arXiv 2408.13656Localize-and-Stitch: Efficient Model Merging via Sparse Task Arithmetic
Yifei He, Yuzheng Hu, Yong Lin et al. · Machine learning
Model merging offers an effective strategy to combine the strengths of multiple finetuned models into a unified model that preserves the specialized capabilities of each.
- 2024-08arXiv 2408.12076ConflictBank: A Benchmark for Evaluating the Influence of Knowledge Conflicts in LLM
Zhaochen Su, Jun Zhang, Xiaoye Qu et al. · Language models & NLP
Large language models (LLMs) have achieved impressive advancements across numerous disciplines, yet the critical issue of knowledge conflicts, a major source of hallucinations, has rarely been studied.
- 2024-08arXiv 2408.07666Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities
Enneng Yang, Li Shen, Guibing Guo et al. · Machine learning
Model merging is an efficient empowerment technique in the machine learning community that does not require the collection of raw training data and does not require expensive computation.
- 2024-08arXiv 2408.06292The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery
Chris Lu, Cong Lu, Robert Tjarko Lange et al. · AI & agents
One of the grand challenges of artificial general intelligence is developing agents capable of conducting scientific research and discovering new knowledge. for brainstorming ideas, writing code, or prediction tasks, they still conduct only a small part of the scientific process.
- 2024-08arXiv 2408.03314Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
Charlie Snell, Jaehoon Lee, Kelvin Xu et al. · Machine learning
Enabling LLMs to improve their outputs by using more test-time computation is a critical step towards building generally self-improving agents that can operate on open-ended natural language.
- 2024-08arXiv 2408.00761Tamper-Resistant Safeguards for Open-Weight LLMs
Rishub Tamirisa, Bhrugu Bharathi, Long Phan et al. · Machine learning
Rapid advances in the capabilities of large language models (LLMs) have raised widespread concerns regarding their potential for malicious use. Open-weight LLMs present unique challenges, as existing safeguards lack robustness to tampering attacks that modify model weights.
- 2024-07arXiv 2407.21770MoMa: Efficient Early-Fusion Pre-training with Mixture of Modality-Aware Experts
Xi Victoria Lin, Akshat Shrivastava, Liang Luo et al. · AI & agents
We introduce MoMa, a novel modality-aware mixture-of-experts (MoE) architecture designed for pre-training mixed-modal, early-fusion language models. MoMa processes images and text in arbitrary sequences by dividing expert modules into modality-specific groups.
- 2024-07arXiv 2407.21787Large Language Monkeys: Scaling Inference Compute with Repeated Sampling
Bradley Brown, Jordan Juravsky, Ryan Ehrlich et al. · Machine learning
Scaling the amount of compute used to train language models has dramatically improved their capabilities. However, when it comes to inference, we often limit models to making only one attempt at a problem.
- 2024-07arXiv 2407.19985Mixture of Nested Experts: Adaptive Processing of Visual Tokens
Gagan Jain, Nidhi Hegde, Aditya Kusupati et al. · Vision
The visual medium (images and videos) naturally contains a large amount of information redundancy, thereby providing a great opportunity for leveraging efficiency in processing.
- 2024-07arXiv 2407.18581Dynamic Language Group-Based MoE: Enhancing Code-Switching Speech Recognition with Hierarchical Routing
Hukai Huang, Shenghui Lu, Yahui Shan et al. · Language models & NLP
The Mixture of Experts (MoE) model is a promising approach for handling code-switching speech recognition (CS-ASR) tasks. However, the existing CS-ASR work on MoE has yet to leverage the advantages of MoE's parameter scaling ability fully.
- 2024-07arXiv 2407.18901AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents
Harsh Trivedi, Tushar Khot, Mareike Hartmann et al. · Society, software & economics
, notes, messaging, shopping app) via APIs, but also generate rich code with complex control flow in an iterative manner based on their interaction with the environment.
Cited inInternal notes - 2024-07arXiv 2407.18418Know Your Limits: A Survey of Abstention in Large Language Models
Bingbing Wen, Jihan Yao, Shangbin Feng et al. · Language models & NLP
Abstention, the refusal of large language models (LLMs) to provide an answer, is increasingly recognized for its potential to mitigate hallucinations and enhance safety in LLM systems.
- 2024-07arXiv 2407.21057Multi-group Uncertainty Quantification for Long-form Text Generation
Terrance Liu, Zhiwei Steven Wu · Language models & NLP
While past works have shown how uncertainty quantification can be applied to large language model (LLM) outputs, the question of whether resulting uncertainty guarantees still hold within sub-groupings of data remains open.
- 2024-07arXiv 2407.12404Analyzing the Generalization and Reliability of Steering Vectors
Daniel Tan, David Chanin, Aengus Lynch et al. · Machine learning
Steering vectors (SVs) have been proposed as an effective approach to adjust language model behaviour at inference time by intervening on intermediate model activations.
- 2024-07arXiv 2407.12784AgentPoison: Red-teaming LLM Agents via Poisoning Memory or Knowledge Bases
Zhaorun Chen, Zhen Xiang, Chaowei Xiao et al. · Machine learning
LLM agents have demonstrated remarkable performance across various applications, primarily due to their advanced capabilities in reasoning, utilizing external knowledge and tools, calling APIs, and executing actions to interact with environments.
- 2024-07arXiv 2407.10490Learning Dynamics of LLM Finetuning
Yi Ren, Danica J. Sutherland · Machine learning
Learning dynamics, which describes how the learning of specific training examples influences the model's predictions on other examples, gives us a powerful tool for understanding the behavior of deep learning systems.
- 2024-07arXiv 2407.10671Qwen2 Technical Report
An Yang, Baosong Yang, Binyuan Hui et al. · Language models & NLP
This report introduces the Qwen2 series, the latest addition to our large language models and large multimodal models.5 to 72 billion, featuring dense models and a Mixture-of-Experts model.
- 2024-07arXiv 2407.09816MaskMoE: Boosting Token-Level Learning via Routing Mask in Mixture-of-Experts
Zhenpeng Su, Zijia Lin, Xue Bai et al. · Language models & NLP
Scaling the size of a model enhances its capabilities but significantly increases computation complexity. Mixture-of-Experts models (MoE) address the issue by allowing model size to scale up without substantially increasing training or inference costs.
- 2024-07arXiv 2407.09435MUSCLE: A Model Update Strategy for Compatible LLM Evolution
Jessica Echterhoff, Fartash Faghri, Raviteja Vemulapalli et al. · AI & agents
Large Language Models (LLMs) are regularly updated to enhance performance, typically through changes in data or architecture. Within the update process, developers often prioritize improving overall performance metrics, paying less attention to maintaining compatibility with earlier model versions.
- 2024-07arXiv 2407.08351AutoBencher: Towards Declarative Benchmark Construction
Xiang Lisa Li, Farzaan Kaiyom, Evan Zheran Liu et al. · Language models & NLP
We present AutoBencher, a declarative framework for automatic benchmark construction, and use it to scalably discover novel insights and vulnerabilities of existing language models.
- 2024-07arXiv 2407.08699Mitigating Catastrophic Forgetting in Language Transfer via Model Merging
Anton Alexandrov, Veselin Raychev, Mark Niklas Müller et al. · Machine learning
As open-weight large language models (LLMs) achieve ever more impressive performances across a wide range of tasks in English, practitioners aim to adapt these models to different languages.
- 2024-07arXiv 2407.06322MagMax: Leveraging Model Merging for Seamless Continual Learning
Daniel Marczak, Bartłomiej Twardowski, Tomasz Trzciński et al. · Machine learning
This paper introduces a continual learning approach named MagMax, which utilizes model merging to enable large pre-trained models to continuously learn from new data without forgetting previously acquired knowledge.
- 2024-07arXiv 2407.04202Reverse Engineering the Fly Brain Using FlyCircuit Database
Yu-Tai Ching, Chin-Ping Cho, Fu-Kai Tang et al. · Neuroscience & neural computation
A method to reverse engineering of a fly brain using the {\it FlyCircuit} database is presented. This method was designed based on the assumption that similar neurons could serve identical functions.
- 2024-07arXiv 2407.04503When LLMs Play the Telephone Game: Cultural Attractors as Conceptual Tools to Evaluate LLMs in Multi-turn Settings
Jérémy Perez, Grgur Kovač, Corentin Léger et al. · Hardware, photonics & physics
As large language models (LLMs) start interacting with each other and generating an increasing amount of text online, it becomes crucial to better understand how information is transformed as it passes from one LLM to the next.
- 2024-07arXiv 2407.04622On scalable oversight with weak LLMs judging strong LLMs
Zachary Kenton, Noah Y. Siegel, János Kramár et al. · Machine learning
Scalable oversight protocols aim to enable humans to accurately supervise superhuman AI.
- 2024-07arXiv 2407.04153Mixture of A Million Experts
Xu Owen He · Machine learning
The feedforward (FFW) layers in standard transformer architectures incur a linear increase in computational costs and activation memory as the hidden layer width grows.
- 2024-07arXiv 2407.12831Truth is Universal: Robust Detection of Lies in LLMs
Lennart Bürger, Fred A. Hamprecht, Boaz Nadler · Language models & NLP
Large Language Models (LLMs) have revolutionised natural language processing, exhibiting impressive human-like capabilities. In particular, LLMs are capable of "lying", knowingly outputting false statements.
Cited inCoherence is selected, anchored, then pruned: a literature synthesis on logical coherence in LLM world modelsStable wrong beliefs evade every self-check: a literature synthesis on model self-knowledge and confident errorsStudy: Logical coherence of LLM world modelsStudy: Model self knowledge and confident errors - 2024-07arXiv 2407.12828Why Does New Knowledge Create Messy Ripple Effects in LLMs?
Jiaxin Qin, Zixuan Zhang, Manling Li et al. · Language models & NLP
Extensive previous research has focused on post-training knowledge editing (KE) for language models (LMs) to ensure that knowledge remains accurate and up-to-date.
Cited inA gradient kernel flags contradictions but does not predict forgetting, and steering rehearsal by it backfiresCoherence is selected, anchored, then pruned: a literature synthesis on logical coherence in LLM world modelsEWC protects what its Fisher sees: a literature synthesis on elastic weight consolidation, scale and coherenceStudy: Elastic weight consolidation deep diveStudy: Logical coherence of LLM world models
Built from every arXiv paper cited in our papers, posts, topic hubs, reports and research notes, with titles, authors, dates and abstracts from the arXiv API (updated 2026-11-12). Summaries are the opening sentences of each paper's own abstract, not our assessment; a citation means we used or checked the work, not that we endorse it. Fields follow each paper's primary arXiv category. Papers cited only outside arXiv (journals, proceedings) are not listed yet. Entries are generated from our citations, so a mistyped ID in a note would list the wrong paper; we remove those as we find them. Report a wrong entry.
