RACE : Large-scale R e A ding comprehension dataset from examinations

representative citing papers

citing papers explorer

Showing 33 of 33 citing papers.

• Open Datasets in Learning Analytics: Trends, Challenges, and Best PRACTICE cs.CY · 2026-02-19 · accept · none · ref 196

  A survey of 172 open educational datasets from 204 papers across LAK, EDM, and AIED conferences reveals trends, 143 previously uncatalogued datasets, field gaps, and an 8-item PRACTICE checklist for better data publication.

• MultiHashFormer: Hash-based Generative Language Models cs.CL · 2026-06-26 · unverdicted · none · ref 19

  MultiHashFormer enables hash-based autoregression in LMs by encoding tokens as multi-hash signatures, outperforming standard Transformers at 100M-3B scales while keeping parameter count constant for multilingual expansion.

• Queryable LoRA: Instruction-Regularized Routing Over Shared Low-Rank Update Atoms cs.LG · 2026-05-08 · unverdicted · none · ref 9

  Queryable LoRA adds dynamic routing over shared low-rank atoms with attention and language-instruction regularization to make parameter-efficient fine-tuning more adaptive across inputs and layers.

• Are LLM Uncertainty and Correctness Encoded by the Same Features? A Functional Dissociation via Sparse Autoencoders cs.LG · 2026-04-21 · unverdicted · none · ref 44

  Uncertainty and correctness in LLMs are encoded by distinct feature populations, with suppression of confounded features improving accuracy and reducing entropy.

• SimDiff: Depth Pruning via Similarity and Difference cs.AI · 2026-04-21 · unverdicted · none · ref 24

  SimDiff uses similarity and difference metrics to prune LLM layers more effectively than cosine similarity alone, retaining over 91% performance at 25% pruning on LLaMA2-7B.

• Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Models cs.AI · 2024-06-14 · conditional · none · ref 209

  LLMs trained on simple specification gaming generalize to zero-shot reward tampering including rewriting their own reward function.

• DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model cs.CL · 2024-05-07 · unverdicted · none · ref 105

  DeepSeek-V2 delivers top-tier open-source LLM performance using only 21B active parameters by compressing the KV cache 93.3% and cutting training costs 42.5% via MLA and DeepSeekMoE.

• GAIA: a benchmark for General AI Assistants cs.CL · 2023-11-21 · unverdicted · none · ref 112

  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.

• BBQ: A Hand-Built Bias Benchmark for Question Answering cs.CL · 2021-10-15 · accept · none · ref 78

  BBQ is a new benchmark dataset showing that QA models often default to social stereotypes, achieving up to 3.4 points higher accuracy when the correct answer aligns with bias.

• The Power of Scale for Parameter-Efficient Prompt Tuning cs.CL · 2021-04-18 · unverdicted · none · ref 24

  Prompt tuning matches full model tuning performance on large language models while tuning only a small fraction of parameters and improves robustness to domain shifts.

• ALBERT: A Lite BERT for Self-supervised Learning of Language Representations cs.CL · 2019-09-26 · accept · none · ref 22

  ALBERT reduces BERT parameters via embedding factorization and layer sharing, adds inter-sentence coherence pretraining, and reaches SOTA on GLUE, RACE, and SQuAD with fewer parameters than BERT-large.

• BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions cs.CL · 2019-05-24 · accept · none · ref 18

  BoolQ introduces naturally occurring yes/no questions as a challenging benchmark where BERT fine-tuned on MultiNLI reaches 80.4% accuracy against 90% human performance.

• Understanding Layer Patching in Model Size Interpolation cs.LG · 2026-07-09 · conditional · none · ref 26

  Optimal layer-patching order for boomerang distillation is a shortest path on a KL-weighted Boolean lattice; greedy KLPatch and simple sequential orders often yield near-optimal interpolations.

• Surrogate Fidelity: When Can Open LLMs Explain Closed Ones? cs.LG · 2026-06-30 · unverdicted · none · ref 68

  Prediction agreement between open and closed LLMs substantially overstates agreement on attributions and causal reasons.

• XDomainBench: Diagnosing Reasoning Collapse in High-Dimensional Scientific Knowledge Composition cs.AI · 2026-05-14 · unverdicted · none · ref 5

  XDomainBench shows LLMs suffer systematic reasoning collapse as domain composition order increases due to direct difficulty and interaction-amplified failures.

• Attention to Mamba: A Recipe for Cross-Architecture Distillation cs.CL · 2026-04-01 · unverdicted · none · ref 18

  A two-stage distillation recipe converts a Pythia-1B Transformer into a Mamba model that preserves performance with perplexity 14.11 versus the teacher's 13.86.

• SHINE: A Scalable In-Context Hypernetwork for Mapping Context to LoRA in a Single Pass cs.CL · 2026-02-06 · conditional · none · ref 12

  SHINE generates LoRA adapters from a document in one forward pass, letting a frozen LLM answer questions about the document without the document in its context.

• Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoning cs.AI · 2025-07-01 · conditional · none · ref 192

  Math reasoning gains in LLMs rarely transfer to general domains; RL tuning generalizes while SFT causes forgetting and representation drift.

• Mixture-of-Experts Can Surpass Dense LLMs Under Strictly Equal Resource cs.CL · 2025-06-13 · conditional · none · ref 21

  MoE models with activation rates in an optimal region outperform dense LLMs of identical total parameter count, training compute, and data budget, with the optimal region consistent across scales.

• Representation Engineering: A Top-Down Approach to AI Transparency cs.LG · 2023-10-02 · unverdicted · none · ref 2

  Representation engineering uses population-level representations in deep neural networks to monitor and manipulate cognitive phenomena like honesty and harmlessness, providing simple effective baselines for LLM safety.

• Enhancing Chat Language Models by Scaling High-quality Instructional Conversations cs.CL · 2023-05-23 · conditional · none · ref 130

  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.

• BloombergGPT: A Large Language Model for Finance cs.LG · 2023-03-30 · conditional · none · ref 56

  BloombergGPT is a 50B parameter LLM trained on a 708B token mixed financial and general dataset that outperforms prior models on financial benchmarks while preserving general LLM performance.

• SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models cs.CL · 2023-03-15 · unverdicted · none · ref 16

  SelfCheckGPT detects hallucinations by checking consistency across multiple sampled responses from black-box LLMs on WikiBio biography generation tasks.

• Atlas: Few-shot Learning with Retrieval Augmented Language Models cs.CL · 2022-08-05 · unverdicted · none · ref 215

  Atlas reaches over 42% accuracy on Natural Questions with only 64 examples, outperforming a 540B-parameter model by 3% with 50x fewer parameters.

• PaLM: Scaling Language Modeling with Pathways cs.CL · 2022-04-05 · accept · none · ref 80

  PaLM 540B demonstrates continued scaling benefits by setting new few-shot SOTA results on hundreds of benchmarks and outperforming humans on BIG-bench.

• Position: Uncertainty Quantification in LLMs is Just Unsupervised Clustering cs.CL · 2026-05-19 · unverdicted · none · ref 54

  Mainstream UQ for LLMs reduces to unsupervised clustering of internal generation consistency and therefore cannot detect confident hallucinations or provide reliable safety signals.

• SMoE: An Algorithm-System Co-Design for Pushing MoE to the Edge via Expert Substitution cs.AI · 2025-08-26 · unverdicted · none · ref 15

  SMoE substitutes low-importance experts with cached similar ones in MoE inference on edge devices to achieve 48% lower decoding latency and over 60% cache hit rate with nearly lossless accuracy.

• InternLM2 Technical Report cs.CL · 2024-03-26 · unverdicted · none · ref 188

  InternLM2 is a new open-source LLM that outperforms prior versions on 30 benchmarks and long-context tasks through scaled pre-training to 32k tokens and a conditional online RLHF alignment strategy.

• DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models cs.CL · 2024-01-11 · unverdicted · none · ref 31

  DeepSeekMoE 2B matches GShard 2.9B performance and approaches a dense 2B model; the 16B version matches LLaMA2-7B at 40% compute by using fine-grained expert segmentation plus shared experts.

• PaLM 2 Technical Report cs.CL · 2023-05-17 · unverdicted · none · ref 85

  PaLM 2 reports state-of-the-art results on language, reasoning, and multilingual tasks with improved efficiency over PaLM.

• Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale cs.CL · 2026-06-13 · unverdicted · none · ref 114

  Technical report announcing Ling-2.6 and Ring-2.6 models with hybrid linear attention, evolutionary CoT, and KPop RL for efficient agentic intelligence at scale.

• DeepSeek LLM: Scaling Open-Source Language Models with Longtermism cs.CL · 2024-01-05 · unverdicted · none · ref 115

  DeepSeek LLM 67B exceeds LLaMA-2 70B on code, mathematics and reasoning benchmarks after pre-training on 2 trillion tokens and alignment via SFT and DPO.

• A Tree-of-Thoughts Inspired Hybrid Approach for Legal Case Judgement Summarization using LLMs cs.CL · 2026-06-26 · unverdicted · none · ref 266

  A tree-of-thoughts inspired hybrid extractive-abstractive LLM prompt yields better legal case judgment summaries than standard extractive or abstractive prompts.