Title resolution pending

representative citing papers

citing papers explorer

Showing 36 of 36 citing papers.

• ORPO: Monolithic Preference Optimization without Reference Model cs.CL · 2024-03-12 · conditional · none · ref 97

  ORPO performs preference alignment during supervised fine-tuning via a monolithic odds ratio penalty, allowing 7B models to outperform larger state-of-the-art models on alignment benchmarks.

• Beyond Accuracy: Robustness, Interpretability and Expressiveness of EEG Foundation Models cs.LG · 2026-05-17 · unverdicted · none · ref 55

  EEG foundation models show no single winner across failure modes, attend to correct brain regions but decode corrupted signals, and retain task information in early layers while late layers adapt during fine-tuning.

• UCSF-PDGM-VQA: Visual Question Answering dataset for brain tumor MRI interpretation cs.CV · 2026-05-16 · unverdicted · none · ref 34 · 2 links

  Introduces the UCSF-PDGM-VQA dataset of 2387 QA pairs from 473 glioma MRI studies and demonstrates that state-of-the-art VLMs exhibit modality collapse on multi-sequence 3D medical images.

• Language Acquisition Device in Large Language Models cs.CL · 2026-05-16 · unverdicted · none · ref 65

  Pre-pretraining on MP-STRUCT matches k-Shuffle Dyck baselines in efficiency while adding human-like resistance to implausible languages and challenges the need for C-RASP definability in effective PPT languages.

• TokAlign++: Advancing Vocabulary Adaptation via Better Token Alignment cs.CL · 2026-05-13 · unverdicted · none · ref 57

  TokAlign++ learns token alignments between LLM vocabularies from monolingual representations to enable faster adaptation, better text compression, and effective token-level distillation across 15 languages with minimal steps.

• Online Learning-to-Defer with Varying Experts stat.ML · 2026-05-12 · unverdicted · none · ref 87 · 2 links

  Presents the first online Learning-to-Defer algorithm achieving regret O((n + n_e) T^{2/3}) generally and O((n + n_e) sqrt(T)) under low noise for multiclass classification with varying experts.

• AAAC: Activation-Aware Adaptive Codebooks for 4-bit LLM Weight Quantization cs.LG · 2026-05-09 · unverdicted · none · ref 10

  AAAC uses two adaptive 64-byte codebooks per layer for 4-bit LLM weight quantization, choosing the optimal one per group to minimize activation-weighted error with zero storage overhead and fast runtime.

• Taming the Entropy Cliff: Variable Codebook Size Quantization for Autoregressive Visual Generation cs.CV · 2026-05-07 · unverdicted · none · ref 5

  Variable codebook sizes that increase along the sequence in visual tokenizers reduce generation FID scores significantly for autoregressive models on ImageNet.

• DARE: Diffusion Language Model Activation Reuse for Efficient Inference cs.LG · 2026-05-01 · unverdicted · none · ref 6

  DARE reuses up to 87% of attention activations in diffusion LLMs through KV caching and output reuse, delivering 1.2x per-layer latency gains with average performance drops of 1.2-2.0%.

• BWLA: Breaking the Barrier of W1AX Post-Training Quantization for LLMs cs.LG · 2026-05-01 · unverdicted · none · ref 22

  BWLA is the first post-training quantization method for LLMs that achieves 1-bit weights paired with low-bit activations such as 6 bits, using OKT to reshape weights and suppress activation tails plus PSP for low-rank refinement.

• Copy First, Translate Later: Interpreting Translation Dynamics in Multilingual Pretraining cs.CL · 2026-04-19 · unverdicted · none · ref 38

  Multilingual pretraining develops translation in two phases: early copying driven by surface similarities, followed by generalizing mechanisms while copying is refined.

• SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolution cs.SE · 2025-02-25 · unverdicted · none · ref 87

  SWE-RL uses RL on software evolution data to train LLMs achieving 41% on SWE-bench Verified with generalization to other reasoning tasks.

• Your Absorbing Discrete Diffusion Secretly Models the Conditional Distributions of Clean Data cs.LG · 2024-06-06 · conditional · none · ref 15

  Absorbing discrete diffusion models the conditional distributions of clean data; reparameterizing yields a time-independent RADD that unifies with AO-ARMs and reaches SOTA perplexity among diffusion models on zero-shot language benchmarks.

• Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations cs.LG · 2024-02-27 · unverdicted · none · ref 27

  HSTU-based generative recommenders with 1.5 trillion parameters scale as a power law with compute up to GPT-3 scale, outperform baselines by up to 65.8% NDCG, run 5-15x faster than FlashAttention2 on long sequences, and improve online A/B metrics by 12.4%.

• Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levels cs.CV · 2023-12-28 · conditional · none · ref 241

  Q-Align trains LMMs on discrete text-defined levels for visual scoring, achieving SOTA on IQA, IAA, and VQA while unifying the tasks in OneAlign.

• GAIA: a benchmark for General AI Assistants cs.CL · 2023-11-21 · unverdicted · none · ref 35

  GAIA benchmark shows humans at 92% accuracy on simple real-world questions far outperform current AI systems at 15%, proposing this gap as a key milestone for general AI.

• Eliciting Latent Predictions from Transformers with the Tuned Lens cs.LG · 2023-03-14 · accept · none · ref 101

  Training per-layer affine probes on frozen transformers yields more reliable latent predictions than the logit lens and enables detection of malicious inputs from prediction trajectories.

• MixRea: Benchmarking Explicit-Implicit Reasoning in Large Language Models cs.CL · 2026-05-19 · unverdicted · none · ref 53

  MixRea benchmark reveals LLMs achieve at most 42.8% consistency on explicit-implicit reasoning tasks, with PRCP prompting proposed to recover overlooked relations.

• GSM-SEM: Benchmark and Framework for Generating Semantically Variant Augmentations cs.CL · 2026-05-08 · unverdicted · none · ref 65

  GSM-SEM generates reusable, stochastic semantic variants of math reasoning benchmarks that alter underlying facts but preserve answers, producing larger LLM performance drops than prior surface-level variants.

• InfoLaw: Information Scaling Laws for Large Language Models with Quality-Weighted Mixture Data and Repetition cs.CL · 2026-05-04 · unverdicted · none · ref 5

  InfoLaw models pretraining as information accumulation where quality sets information density and repetition causes scale-dependent diminishing returns, predicting loss with low error on unseen mixtures and larger scales up to 7B models and 425B tokens.

• GRASPrune: Global Gating for Budgeted Structured Pruning of Large Language Models cs.AI · 2026-04-21 · unverdicted · none · ref 50

  GRASPrune removes 50% of parameters from LLaMA-2-7B via global gating and projected straight-through estimation, reaching 12.18 WikiText-2 perplexity and competitive zero-shot accuracy after four epochs on 512 calibration sequences.

• Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference cs.CL · 2024-06-16 · unverdicted · none · ref 41

  Quest speeds up long-context LLM self-attention by up to 2.23x via query-dependent selection of top-K critical KV cache pages, cutting overall latency by 7.03x with negligible accuracy loss.

• Back to Basics: Revisiting REINFORCE Style Optimization for Learning from Human Feedback in LLMs cs.LG · 2024-02-22 · conditional · none · ref 83

  REINFORCE-style variants outperform PPO, DPO, and RAFT in RLHF for LLMs by removing unnecessary PPO components and adapting the simpler method to LLM alignment characteristics.

• The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets cs.AI · 2023-10-10 · unverdicted · none · ref 11

  At sufficient scale, LLMs linearly represent the truth value of factual statements, as shown by visualizations, cross-dataset generalization, and causal interventions that flip truth judgments.

• Scaling Relationship on Learning Mathematical Reasoning with Large Language Models cs.CL · 2023-08-03 · unverdicted · none · ref 40

  Pre-training loss predicts LLM math reasoning better than parameter count; rejection sampling fine-tuning with diverse paths raises LLaMA-7B accuracy on GSM8K from 35.9% with SFT to 49.3%.

• Enhancing Chat Language Models by Scaling High-quality Instructional Conversations cs.CL · 2023-05-23 · conditional · none · ref 223

  UltraChat supplies 1.5 million high-quality multi-turn dialogues that, when used to fine-tune LLaMA, produce UltraLLaMA, which outperforms prior open-source chat models including Vicuna.

• CLIC: Contextual Language-Informed Cardiac Pathology Classification cs.LG · 2026-05-18 · unverdicted · none · ref 10

  CLIC encodes patient context and technical metadata as natural language text to boost ECG-based cardiac pathology classification performance over signal-only models.

• From Codebooks to VLMs: Evaluating Automated Visual Discourse Analysis for Climate Change on Social Media cs.CV · 2026-04-23 · unverdicted · none · ref 65

  VLMs recover reliable population-level trends in climate change visual discourse on social media even when per-image accuracy is only moderate.

• GS-Quant: Granular Semantic and Generative Structural Quantization for Knowledge Graph Completion cs.AI · 2026-04-23 · unverdicted · none · ref 137

  GS-Quant generates coarse-to-fine discrete codes for KG entities via semantic hierarchy injection and causal sequence reconstruction, enabling LLMs to perform knowledge graph completion by treating the codes as vocabulary tokens.

• SmolLM2: When Smol Goes Big -- Data-Centric Training of a Small Language Model cs.CL · 2025-02-04 · unverdicted · none · ref 27

  SmolLM2 is a 1.7B-parameter language model that outperforms Qwen2.5-1.5B and Llama3.2-1B after overtraining on 11 trillion tokens using custom FineMath, Stack-Edu, and SmolTalk datasets in a multi-stage pipeline.

• AppAgent: Multimodal Agents as Smartphone Users cs.CV · 2023-12-21 · unverdicted · none · ref 55

  AppAgent lets large language models operate diverse smartphone apps via visual interactions and learns app usage from exploration or demonstrations.

• Gemma: Open Models Based on Gemini Research and Technology cs.CL · 2024-03-13 · accept · none · ref 81

  Gemma introduces open 2B and 7B LLMs derived from Gemini technology that beat comparable open models on 11 of 18 text tasks and come with safety assessments.

• Fine-Tuning Pre-Trained Code Models for AI-Generated Code Detection cs.CL · 2026-05-02 · unverdicted · none · ref 11

  Fine-tuning CodeBERT, GraphCodeBERT, UniXcoder and CodeT5+ with augmentation, cross-validation and ensembling yields macro-F1 of 0.737 on binary human-vs-AI code detection and 0.422 on 11-class model attribution in SemEval-2026 Task 13.

• Gemma 2: Improving Open Language Models at a Practical Size cs.CL · 2024-07-31 · conditional · none · ref 90

  Gemma 2 models achieve leading performance at their sizes by combining established Transformer modifications with knowledge distillation for the 2B and 9B variants.

• From Pixels to Prompts: Vision-Language Models cs.AI · 2026-05-08 · unverdicted · none · ref 8

  An explanatory book that supplies a clear mental map and intuition for how Vision-Language Models combine vision and language capabilities.

• PipeSD: An Efficient Cloud-Edge Collaborative Pipeline Inference Framework with Speculative Decoding cs.DC · 2026-05-13 · unreviewed · ref 62 · 2 links