New England NLP Meeting Series

Accepted Work

Oral (3:30-4:20 PM)

  1. AI Alignment at Your Discretion
    Presenter: Hadi Khalaf
    (Best Paper Awardee! 🏆🎉🎉)
  2. CaKE: Circuit-aware Editing Enables Generalizable Knowledge Learners
    Presenter: Yunzhi Yao
  3. Training Language Models on Synthetic Edit Sequences Improves Code Synthesis
    Presenter: Rob Fergus
    (Outstanding Paper Awardee! 🏆)
  4. Language Models use Lookbacks to Track Beliefs
    Presenter: Nikhil Prakash
  5. How Post-Training Reshapes LLMs: A Mechanistic View on Knowledge, Truthfulness, Refusal, and Confidence
    Presenter: Shichang Zhang
    (Outstanding Paper Awardee! 🏆)

Poster (12:50-2:10 PM)

Listed by random order.

  1. Bayesian Teaching Enables Probabilistic Reasoning in Large Language Models
    Presenter: Linlu Qiu
  2. What's Producible May Not Be Reachable: Measuring the Steerability of Generative Models
    Presenter: Sarah Bentley
  3. Who Evaluates the Evaluations? Objectively Scoring Text-to-Image Prompt Coherence Metrics with (TS2)
    Presenter: Mahsa Khoshnoodi
  4. TOMATO: Assessing Visual Temporal Reasoning Capabilities in Multimodal Foundation Models
    Presenter: Ziyao Shangguan
  5. A LSTM language model learns Hindi-Urdu case-agreement interactions, and has a linear encoding of case
    Presenter: Satoru Ozaki
  6. A Probabilistic Inference Approach to LLM Inference-Time Scaling
    Presenter: Isha Puri
  7. A Taxonomy of Transcendence
    Presenter: Natalie Abreu
  8. Are Foundation Models Foundational? Synthetic Tasks Reveal World Models
    Presenter: Peter Chang
  9. Between Circuits and Chomsky: Pre-pretraining on Formal Languages Imparts Linguistic Biases
    Presenter: William Merrill
  10. Boosting Large Language Models with Mask Fine-Tuning
    Presenter: Yue Bai
  11. Building A Unified AI-centric Language System: analysis, framework and future work
    Presenter: Edward Hong Wang
  12. Can model interpretations predict behavior on unseen data?
    Presenter: Victoria Li
  13. Capturing Human Cognitive Styles with Language: Towards an Experimental Evaluation Paradigm
    Presenter: Vasudha Varadarajan
  14. ChatGPT Doesn't Trust Chargers Fans: Guardrail Sensitivity in Context
    Presenter: Victoria Li
  15. Chunk-Distilled Language Modeling
    Presenter: Jiawei (Joe) Zhou
  16. Classical Computation in Connectionist Models
    Presenter: Aditya Yedetore
  17. CodeSteer: Symbolic-Augmented Language Models via Code/Text Guidance
    Presenter: Yongchao Chen
  18. Communication Makes Perfect: Persuasion Dataset Construction via Multi-LLM Communication
    Presenter: Weicheng Ma
  19. Contextual morphologically-guided tokenization for pretrained Latin BERT models
    Presenter: Marisa Hudspeth
  20. Continued Pre-training LLMs to Learn Simulated Knowledge Updates
    Presenter: Aochong Oliver Li
  21. Do Automatic Factuality Metrics Measure Factuality?
    Presenter: Sanjana Ramprasad
  22. Escaping Collapse: The Strength of Weak Data for Large Language Model Training
    Presenter: Alex Bie
  23. Explaining GPT-4's Schema of Depression Using Machine Behavior Analysis
    Presenter: Adithya V Ganesan
  24. Faithful, Unfaithful or Ambiguous? Multi-Agent Debate with Initial Stance for Summary Evaluation
    Presenter: Mahnaz Koupaee
  25. Fast Controlled Generation from Language Models with Adaptive Weighted Rejection Sampling
    Presenter: Ben Lipkin
  26. Focus Directions Make Your Language Models Pay More Attention to Relevant Contexts
    Presenter: Youxiang Zhu
  27. Generating Text from Uniform Meaning Representation
    Presenter: Emma Markle
  28. HYBRIDMIND: Meta Selection of Natural Language and Symbolic Language for Enhanced LLM Reasoning
    Presenter: Simeng Han
  29. In Search of Lost Language Model Training Dynamics
    Presenter: Zhenting Qi
  30. In-Context Learning of Representations
    Presenter: Ekdeep Singh Lubana
  31. Inductive Linguistic Reasoning with Large Language Models
    Presenter: Raghav Ramji
  32. Is analogy enough to draw novel adjective-noun inferences?
    Presenter: Hayley Ross
  33. JumpStarter: A Multi-Agent System for Getting Started on Personal Goals via Adaptive Personal Context Curation
    Presenter: Xuanming Zhang
  34. K-Paths: Reasoning over Graph Paths for Drug Repurposing and Drug Interaction Prediction
    Presenter: Tassallah Amina Abdullahi
  35. KARL: Knowledge-Aware Retrieval and Representations aid Retention and Learning in Students
    Presenter: Matthew Shu
  36. LLMs can Perform Multi-Dimensional Analytic Writing Assessments: A Case Study of L2 Graduate-Level Academic English Writing
    Presenter: Zhengxiang Wang
  37. Loss in the Crowd: Hidden Breakthroughs in LM Training
    Presenter: Sara Kangaslahti
  38. Mind the Gap: Assessing Crowd-Sourced Linguistic Knowledge on Morphological Gaps of Two Related Languages
    Presenter: Jonathan Sakunkoo
  39. Name of Thrones: Evaluating How LLMs Rank Student Names, Race, and Gender in Status Hierarchies
    Presenter: Annabella Sakunkoo
  40. NüshuRescue: Reviving the Endangered Nüshu Language with AI
    Presenter: Ivory Yang
  41. Performing Scientific Research with Artificial Intelligence Researcher: A Comprehensive Study with Expert-Involved Evaluation
    Presenter: Tianyu Liu
  42. Planetarium🪐: A Rigorous Benchmark for Translating Text to Structured Planning Languages
    Presenter: Max Zuo
  43. Potemkin Understanding in Large Language Models: Formalizing and Benchmarking Conceptual Comprehension
    Presenter: Marina Mancoridis
  44. Privacy Ripple Effects from Adding or Removing Personal Information in Language Model Training
    Presenter: Jaydeep Borkar
  45. Probing the Capacity of Language Model Agents to Operationalize Disparate Experiential Context Despite Distraction
    Presenter: Sonny George
  46. Projecting Assumptions: The Duality Between Sparse Autoencoders and Concept Geometry
    Presenter: Sumedh Hindupur
  47. PUPA: Private User Prompt Annotations Benchmark
    Presenter: Siyan Li
  48. Re-Evaluating Evaluation for Multilingual Summarization
    Presenter: Jessica Forde
  49. Scaling Makes It Possible: How Large Models Master Impossible Languages
    Presenter: Yaning Jia
  50. Scaling Sparse and Dense Retrieval in Decoder-Only LLMs
    Presenter: Hansi Zeng
  51. Scaling the Wall: Scaling Vanilla RNNs by Stealing Transformer Geometry
    Presenter: Vighnesh Subramaniam
  52. Self-Steering Language Models
    Presenter: Gabriel Grand
  53. Sociolinguistic Simulacra: Interactions Between Language and Attitudes in Finetuned Language Models
    Presenter: Carter Teplica
  54. Steering Fine Tuning with Targeted Concept Ablation
    Presenter: Caden Juang
  55. Superpower🦸⚡️ of the Contrastive Decoding📈 comes from its Imagination🧠💡!
    Presenter: Haw-Shiuan Chang
  56. Syntactic and Semantic Control of Large Language Models via Sequential Monte Carlo
    Presenter: João Loula
  57. Temporal Working Memory: Query-Guided Segment Refinement for Enhanced Multimodal Understanding
    Presenter: Xingjian Diao
  58. TextArena: Beyond Traditional Benchmarks - Evaluating Social Intelligence in Language Models
    Presenter: Simon Yu
  59. The Same but Different: Structural Similarities and Differences in Multilingual Language Modeling
    Presenter: Ruochen Zhang
  60. The Semantic Hub Hypothesis: Language Models Share Semantic Representations Across Languages and Modalities
    Presenter: Zhaofeng Wu
  61. Toward Reliable Ad-hoc Scientific Information Extraction: A Case Study on Two Materials Datasets
    Presenter: Satanu Ghosh
  62. Working Memory Identifies Reasoning Limits in Language Models
    Presenter: Chunhui Zhang
  63. Auto-encoding Scientific Conclusions for Hypothesis Generation
    Presenter: Brian Ondov
  64. (How) Do Language Models Track State?
    Presenter: Belinda Zou Li
  65. A Systematic Evaluation of Transformer-LM Representations for Capturing Author States and Traits
    Presenter: Khushboo Singh
  66. Accelerating robust in-context language learning
    Presenter: Johnathan Sun
  67. Controlling Factual Associations and Visual Perception in Vision-Language Models
    Presenter: Michal Golovanevsky
  68. Discovering Forbidden Topics in Language Models
    Presenter: Can Rager
  69. Do LLMs synthesize technical information like humans?
    Presenter: Jillian Ross
  70. Early Detection of Mild Cognitive Impairment Through Voice Assistant Interactions: An LLM-Driven Approach
    Presenter: Kristin Qi
  71. Evolutionary Dynamics of Syntax and Semantics in BERT: A Hyperbolic Geometry Perspective
    Presenter: Sukanya Krishna
  72. Exploring the Emergence of Shared Multilingual Concept Representations in LLMs
    Presenter: Kerem Sahin
  73. First things first: Universal path dependence of learning
    Presenter: David Mayo
  74. ImpScore: A Learnable Metric For Quantifying The Implicitness Level of Sentences
    Presenter: Yuxin Wang
  75. Investigating the Knowledge-Perception Trade-off in Vision-Language Models through Visual Counterfactuals
    Presenter: Michal Golovanevsky
  76. Modeling Noisy-Channel Language Processing with Reanalysis of Possible Errors: A Probabilistic Inference Approach
    Presenter: Thomas Clark
  77. Paths Not Taken: Optimize Multilingual Factual Recall Pathways via Simple Zero-Rank Interventions
    Presenters: Meng Lu, Ruochen Zhang, Ellie Pavlick, Carsten Eickhoff
  78. Reasoning-based Regression: Teaching Language Models to Score Natural Language Features
    Presenter: Diane Tchuindjo
  79. Supporting Biomedical Discovery with Human Agent Collaboration for Literature Grounded Search and Reasoning
    Presenter: Shannon Shen
  80. The Dual-Route Model of Induction
    Presenter: Sheridan Feucht
  81. The Role of PropBank Sense Numbers in AMR-to-text Generation and Text-to-AMR Parsing
    Presenter: Thu Hoang
  82. ThoughtCoder: Structured and adaptive problem solving via language model programming
    Presenter: Ced Zhang
  83. Reviving Endangered and Extinct Languages with Large Language Models
    Presenter: Weicheng Ma
  84. United We Stand: Multi-LLM Collaboration for Advancing Scientific Research
    Presenter: Weicheng Ma
  85. What's Hidden in Flemish Stories: How Can LLMs Unveil the Affective Nuances in Daily Narratives?
    Presenters: Ratna Kandala, Katie Hoemann