Standardized re-evaluation of CRS methods on ReDial finds that nearly half of reported accuracy stems from repetition shortcuts absent in novelty-focused tests, performance tracks LLM capacity more than architecture, and traditional recall overstates conversational utility.
hub
QLoRA: Efficient Finetuning of Quantized LLMs
48 Pith papers cite this work. Polarity classification is still indexing.
48
Pith papers citing it
abstract
We present QLoRA, an efficient finetuning approach that reduces memory usage enough to finetune a 65B parameter model on a single 48GB GPU while preserving full 16-bit finetuning task performance. QLoRA backpropagates gradients through a frozen, 4-bit quantized pretrained language model into Low Rank Adapters~(LoRA). Our best model family, which we name Guanaco, outperforms all previous openly released models on the Vicuna benchmark, reaching 99.3% of the performance level of ChatGPT while only requiring 24 hours of finetuning on a single GPU. QLoRA introduces a number of innovations to save memory without sacrificing performance: (a) 4-bit NormalFloat (NF4), a new data type that is information theoretically optimal for normally distributed weights (b) double quantization to reduce the average memory footprint by quantizing the quantization constants, and (c) paged optimziers to manage memory spikes. We use QLoRA to finetune more than 1,000 models, providing a detailed analysis of instruction following and chatbot performance across 8 instruction datasets, multiple model types (LLaMA, T5), and model scales that would be infeasible to run with regular finetuning (e.g. 33B and 65B parameter models). Our results show that QLoRA finetuning on a small high-quality dataset leads to state-of-the-art results, even when using smaller models than the previous SoTA. We provide a detailed analysis of chatbot performance based on both human and GPT-4 evaluations showing that GPT-4 evaluations are a cheap and reasonable alternative to human evaluation. Furthermore, we find that current chatbot benchmarks are not trustworthy to accurately evaluate the performance levels of chatbots. A lemon-picked analysis demonstrates where Guanaco fails compared to ChatGPT. We release all of our models and code, including CUDA kernels for 4-bit training.
hub tools
citation-role summary
background 2
citation-polarity summary
claims ledger
- abstract We present QLoRA, an efficient finetuning approach that reduces memory usage enough to finetune a 65B parameter model on a single 48GB GPU while preserving full 16-bit finetuning task performance. QLoRA backpropagates gradients through a frozen, 4-bit quantized pretrained language model into Low Rank Adapters~(LoRA). Our best model family, which we name Guanaco, outperforms all previous openly released models on the Vicuna benchmark, reaching 99.3% of the performance level of ChatGPT while only requiring 24 hours of finetuning on a single GPU. QLoRA introduces a number of innovations to save m
co-cited works
roles
background 2polarities
background 2representative citing papers
DiagnosticIQ benchmark shows frontier LLMs perform similarly on standard rule-to-action tasks but lose substantial accuracy under distractor expansion and condition inversion, pointing to calibration as the key deployment issue.
Presents MBFC-2025 dataset and multi-view embeddings with fusion methods for media bias and factuality, reporting SOTA results on ACL-2020 and new benchmarks on MBFC-2025.
VLA models exhibit a compute-bound VLM phase followed by a memory-bound action phase on edge hardware; DP-Cache and V-AEFusion reduce redundancy and enable pipeline parallelism for up to 6x speedup on NPUs with marginal task degradation.
Visual token pruning in MLLMs fails on complex reasoning due to Relevant Visual Information Shift during decoding, but the DSTP framework fixes it training-free across models.
CrashSight is a new infrastructure-focused benchmark showing that state-of-the-art vision-language models can describe crash scenes but fail at temporal and causal reasoning.
AtlasOCR delivers the first open-source Darija OCR by fine-tuning Qwen2.5-VL 3B, achieving state-of-the-art results on custom and existing benchmarks for both Darija and Arabic.
KITE is a training-free method that uses keyframe-indexed tokenized evidence including BEV schematics to enhance VLM performance on robot failure detection, identification, localization, explanation, and correction.
Medusa augments LLMs with multiple decoding heads and tree-based attention to predict and verify several tokens in parallel, yielding 2.2-3.6x inference speedup via two fine-tuning regimes.
Iterative self-rewarding via LLM-as-Judge in DPO training on Llama 2 70B improves instruction following and self-evaluation, outperforming GPT-4 on AlpacaEval 2.0.
MinT enables efficient management of million-scale LoRA-adapted LLM policies over shared 1T-parameter base models by moving only small adapters through training and serving pipelines.
Summing outputs from separately trained QLoRA PEFT modules provides strong performance for attribute-controlled text generation, often matching or exceeding single-task modules even on single-attribute tests.
Sharpness-aware pretraining and related flat-minima interventions reduce catastrophic forgetting by up to 80% after post-training across 20M-150M models and by 31-40% at 1B scale.
SST V2 introduces parallel-trainable nonlinear recurrence in latent space to let transformers reason continuously across positions, delivering +15 points on GPQA-Diamond and halving remaining GSM8K errors over matched baselines.
TOFU loss mitigates the narrowing of generative diversity in LLMs after supervised fine-tuning by addressing neglect of low-frequency patterns and forgetting of prior knowledge.
Fine-tuning 7B code LLMs on a custom multi-file DSL dataset achieves structural fidelity of 1.00, high exact-match accuracy, and practical utility validated by expert survey and execution checks.
A separable expert architecture uses base models, LoRA adapters, and deletable per-user proxies to enable privacy-preserving personalization and deterministic unlearning in LLMs.
Pioneer Agent automates the full lifecycle of adapting and continually improving small language models via diagnosis-driven data synthesis and regression-constrained retraining, delivering gains of 1.6-83.8 points on benchmarks and large lifts in production-style tasks.
In Llama 3.1 8B, task-sensitive layers cluster late while RoPE adaptation is strongest early, yet applying both adaptations only to sensitivity-identified layers outperforms other layer choices by 4-16 points on MMLU, GPQA, HumanEval+, MATH, MGSM and ARC.
ORPO is most effective at misaligning LLMs while DPO excels at realigning them, though it reduces utility, revealing an asymmetry between attack and defense methods.
ForkKV uses copy-on-write disaggregated KV cache with DualRadixTree and ResidualAttention kernels to deliver up to 3x throughput over prior multi-LoRA serving systems with negligible quality loss.
CDWF achieves 90-99% of full fine-tuning performance with up to 120x fewer trainable parameters by dynamically allocating full trainability to gradient-important blocks and LoRA to others for PV cyberattack transfer learning.
Gradient-guided layer selection for LoRA yields 15-28% training speedup with matched downstream results on MMLU, GSM8K, and HumanEval across 14 models from 0.5B to 72B parameters.
A VLM framework with spatial patch cross-attention and adaptive PID-Tversky loss reports 90.69% classification accuracy, 0.9512 Dice score, and 92.80 CIDEr for LSS diagnosis plus automated report generation.
citing papers explorer
-
A Standardized Re-evaluation of Conversational Recommender Systems on the ReDial Dataset
Standardized re-evaluation of CRS methods on ReDial finds that nearly half of reported accuracy stems from repetition shortcuts absent in novelty-focused tests, performance tracks LLM capacity more than architecture, and traditional recall overstates conversational utility.
-
DiagnosticIQ: A Benchmark for LLM-Based Industrial Maintenance Action Recommendation from Symbolic Rules
DiagnosticIQ benchmark shows frontier LLMs perform similarly on standard rule-to-action tasks but lose substantial accuracy under distractor expansion and condition inversion, pointing to calibration as the key deployment issue.
-
A Multi-View Media Profiling Suite: Resources, Evaluation, and Analysis
Presents MBFC-2025 dataset and multi-view embeddings with fusion methods for media bias and factuality, reporting SOTA results on ACL-2020 and new benchmarks on MBFC-2025.
-
Characterizing Vision-Language-Action Models across XPUs: Constraints and Acceleration for On-Robot Deployment
VLA models exhibit a compute-bound VLM phase followed by a memory-bound action phase on edge hardware; DP-Cache and V-AEFusion reduce redundancy and enable pipeline parallelism for up to 6x speedup on NPUs with marginal task degradation.
-
Why and When Visual Token Pruning Fails? A Study on Relevant Visual Information Shift in MLLMs Decoding
Visual token pruning in MLLMs fails on complex reasoning due to Relevant Visual Information Shift during decoding, but the DSTP framework fixes it training-free across models.
-
CrashSight: A Phase-Aware, Infrastructure-Centric Video Benchmark for Traffic Crash Scene Understanding and Reasoning
CrashSight is a new infrastructure-focused benchmark showing that state-of-the-art vision-language models can describe crash scenes but fail at temporal and causal reasoning.
-
AtlasOCR: Building the First Open-Source Darija OCR Model with Vision Language Models
AtlasOCR delivers the first open-source Darija OCR by fine-tuning Qwen2.5-VL 3B, achieving state-of-the-art results on custom and existing benchmarks for both Darija and Arabic.
-
KITE: Keyframe-Indexed Tokenized Evidence for VLM-Based Robot Failure Analysis
KITE is a training-free method that uses keyframe-indexed tokenized evidence including BEV schematics to enhance VLM performance on robot failure detection, identification, localization, explanation, and correction.
-
Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads
Medusa augments LLMs with multiple decoding heads and tree-based attention to predict and verify several tokens in parallel, yielding 2.2-3.6x inference speedup via two fine-tuning regimes.
-
Self-Rewarding Language Models
Iterative self-rewarding via LLM-as-Judge in DPO training on Llama 2 70B improves instruction following and self-evaluation, outperforming GPT-4 on AlpacaEval 2.0.
-
MinT: Managed Infrastructure for Training and Serving Millions of LLMs
MinT enables efficient management of million-scale LoRA-adapted LLM policies over shared 1T-parameter base models by moving only small adapters through training and serving pipelines.
-
Output Composability of QLoRA PEFT Modules for Plug-and-Play Attribute-Controlled Text Generation
Summing outputs from separately trained QLoRA PEFT modules provides strong performance for attribute-controlled text generation, often matching or exceeding single-task modules even on single-attribute tests.
-
Sharpness-Aware Pretraining Mitigates Catastrophic Forgetting
Sharpness-aware pretraining and related flat-minima interventions reduce catastrophic forgetting by up to 80% after post-training across 20M-150M models and by 31-40% at 1B scale.
-
State Stream Transformer (SST) V2: Parallel Training of Nonlinear Recurrence for Latent Space Reasoning
SST V2 introduces parallel-trainable nonlinear recurrence in latent space to let transformers reason continuously across positions, delivering +15 points on GPQA-Diamond and halving remaining GSM8K errors over matched baselines.
-
Diversity in Large Language Models under Supervised Fine-Tuning
TOFU loss mitigates the narrowing of generative diversity in LLMs after supervised fine-tuning by addressing neglect of low-frequency patterns and forgetting of prior knowledge.
-
Leveraging LLMs for Multi-File DSL Code Generation: An Industrial Case Study
Fine-tuning 7B code LLMs on a custom multi-file DSL dataset achieves structural fidelity of 1.00, high exact-match accuracy, and practical utility validated by expert survey and execution checks.
-
Separable Expert Architecture: Toward Privacy-Preserving LLM Personalization via Composable Adapters and Deletable User Proxies
A separable expert architecture uses base models, LoRA adapters, and deletable per-user proxies to enable privacy-preserving personalization and deterministic unlearning in LLMs.
-
Pioneer Agent: Continual Improvement of Small Language Models in Production
Pioneer Agent automates the full lifecycle of adapting and continually improving small language models via diagnosis-driven data synthesis and regression-constrained retraining, delivering gains of 1.6-83.8 points on benchmarks and large lifts in production-style tasks.
-
Sensitivity-Positional Co-Localization in GQA Transformers
In Llama 3.1 8B, task-sensitive layers cluster late while RoPE adaptation is strongest early, yet applying both adaptations only to sensitivity-identified layers outperforms other layer choices by 4-16 points on MMLU, GPQA, HumanEval+, MATH, MGSM and ARC.
-
The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training
ORPO is most effective at misaligning LLMs while DPO excels at realigning them, though it reduces utility, revealing an asymmetry between attack and defense methods.
-
ForkKV: Scaling Multi-LoRA Agent Serving via Copy-on-Write Disaggregated KV Cache
ForkKV uses copy-on-write disaggregated KV cache with DualRadixTree and ResidualAttention kernels to deliver up to 3x throughput over prior multi-LoRA serving systems with negligible quality loss.
-
Constraint-Driven Warm-Freeze for Efficient Transfer Learning in Photovoltaic Systems
CDWF achieves 90-99% of full fine-tuning performance with up to 120x fewer trainable parameters by dynamically allocating full trainability to gradient-important blocks and LoRA to others for PV cyberattack transfer learning.
-
Aletheia: Gradient-Guided Layer Selection for Efficient LoRA Fine-Tuning Across Architectures
Gradient-guided layer selection for LoRA yields 15-28% training speedup with matched downstream results on MMLU, GSM8K, and HumanEval across 14 models from 0.5B to 72B parameters.
-
An Explainable Vision-Language Model Framework with Adaptive PID-Tversky Loss for Lumbar Spinal Stenosis Diagnosis
A VLM framework with spatial patch cross-attention and adaptive PID-Tversky loss reports 90.69% classification accuracy, 0.9512 Dice score, and 92.80 CIDEr for LSS diagnosis plus automated report generation.
-
LiFT: Does Instruction Fine-Tuning Improve In-Context Learning for Longitudinal Modelling by Large Language Models?
LiFT instruction fine-tunes LLMs with a temporal curriculum to improve in-context learning on longitudinal NLP tasks, yielding gains on out-of-distribution data and rare change events across multiple model sizes.
-
AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models
AutoDAN automatically generates semantically meaningful jailbreak prompts for aligned LLMs via a hierarchical genetic algorithm, achieving higher attack success, cross-model transferability, and universality than baselines while bypassing perplexity defenses.
-
MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Models
Bootstrapping math questions via rewriting creates MetaMathQA; fine-tuning LLaMA-2 on it yields 66.4% on GSM8K for 7B and 82.3% for 70B, beating prior same-size models by large margins.
-
Baseline Defenses for Adversarial Attacks Against Aligned Language Models
Baseline defenses including perplexity-based detection, input preprocessing, and adversarial training offer partial robustness to text adversarial attacks on LLMs, with challenges arising from weak discrete optimizers.
-
Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena
GPT-4 as an LLM judge achieves over 80% agreement with human preferences on MT-Bench and Chatbot Arena, matching human agreement levels and providing a scalable evaluation method.
-
Fine-Tuning Models for Automated Code Review Feedback
PEFT fine-tuning of Code Llama yields feedback on student Java bugs that students judge equal to ChatGPT and better than prompt engineering, using BLEU/ROUGE/BERTScore plus human ratings.
-
Towards a Large Language-Vision Question Answering Model for MSTAR Automatic Target Recognition
A fine-tuned large language-vision model achieves 98% accuracy on visual question answering for military vehicle identification in SAR imagery from an extended MSTAR benchmark.
-
AdaMeZO: Adam-style Zeroth-Order Optimizer for LLM Fine-tuning Without Maintaining the Moments
AdaMeZO adapts Adam moment estimates to zeroth-order LLM fine-tuning without extra memory storage, outperforming MeZO with up to 70% fewer forward passes.
-
ChipLingo: A Systematic Training Framework for Large Language Models in EDA
ChipLingo trains LLMs on EDA data via corpus construction, domain-adaptive pretraining, and RAG scenario alignment, reaching 59.7% accuracy with an 8B model and 70.02% with a 32B model on a new internal EDA benchmark.
-
A KL Lens on Quantization: Fast, Forward-Only Sensitivity for Mixed-Precision SSM-Transformer Models
KL divergence provides a superior forward-only metric for identifying quantization-sensitive parts in SSM-Transformer hybrids, outperforming MSE and SQNR and supporting practical mixed-precision deployment on edge devices.
-
Please Make it Sound like Human: Encoder-Decoder vs. Decoder-Only Transformers for AI-to-Human Text Style Transfer
BART-large outperforms Mistral-7B in AI-to-human style transfer with higher reference similarity scores and far fewer parameters, while showing that marker shift can reflect overshoot rather than accurate transfer.
-
NyayaMind- A Framework for Transparent Legal Reasoning and Judgment Prediction in the Indian Legal System
NyayaMind combines RAG retrieval with domain-specific LLMs to generate transparent, structured legal reasoning and judgment predictions for Indian court cases.
-
PassiveQA: A Three-Action Framework for Epistemically Calibrated Question Answering via Supervised Finetuning
PassiveQA trains models via supervised finetuning to decide Answer, Ask, or Abstain using structured information-state representations and knowledge-graph context, yielding better abstention and lower hallucination on QA datasets.
-
LLiMba: Sardinian on a Single GPU -- Adapting a 3B Language Model to a Vanishing Romance Language
Qwen2.5-3B was continued-pretrained and then fine-tuned with rsLoRA r256 on Sardinian data to reach 28.5 BLEU into the language, outperforming full fine-tuning and other LoRA variants.
-
OpenSOC-AI: Democratizing Security Operations with Parameter Efficient LLM Log Analysis
LoRA fine-tuning of TinyLlama-1.1B on 450 SOC examples produces 68% threat classification accuracy and 58% severity accuracy on 50 held-out logs, with full code, weights, and data released.
-
Toward Zero-Egress Psychiatric AI: On-Device LLM Deployment for Privacy-Preserving Mental Health Decision Support
A cross-platform mobile application deploys an ensemble of quantized open-source LLMs for fully local, DSM-5-aligned psychiatric decision support with claimed accuracy comparable to prior cloud versions.
-
FLeX: Fourier-based Low-rank EXpansion for multilingual transfer
LoRA fine-tuning of Code Llama with Fourier regularization raises Java pass@1 from 34.2% to 42.1% while using a small high-quality dataset.
-
Information Extraction from Electricity Invoices with General-Purpose Large Language Models
Few-shot prompting lifts F1 scores above 96 percent on electricity-invoice extraction for Gemini 1.5 Pro and Mistral-small, while hyperparameter changes produce only marginal gains.
-
Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey
A comprehensive survey of PEFT algorithms for large models, covering their performance, overhead, applications, and real-world system implementations.
-
SLM Finetuning for Natural Language to Domain Specific Code Generation in Production
Fine-tuned small language models outperform larger models in natural language to domain-specific code generation with improved performance, latency, and the ability to adapt to customer-specific scenarios without losing general capabilities.
-
Large Language Models: A Survey
The paper surveys key large language models, their training methods, datasets, evaluation benchmarks, and future research directions in the field.
-
FUTURAL: A Metasearch Platform for Empowering Rural Areas with Smart Solutions
FUTURAL's metasearch MVP uses LLMs to enable natural language queries over smart solutions data to support rural development.
-
The nextAI Solution to the NeurIPS 2023 LLM Efficiency Challenge
A competition entry achieved efficient fine-tuning of LLaMa2 70B on one GPU in 24 hours with competitive QA benchmark performance.
-
QU-NLP at ArchEHR-QA 2026: Two-Stage QLoRA Fine-Tuning of Qwen3-4B for Patient-Oriented Clinical Question Answering and Evidence Sentence Alignment
Two-stage QLoRA fine-tuning of Qwen3-4B plus retrieval ensemble achieves 32.87 overall score on clinical QA and 67.16 micro-F1 on evidence alignment, highlighting that 20 training cases are insufficient.