Image de couverture de livre

Importance Sampling is All You Need: Predict LLM's performance on new benchmark by reusing existing benchmark

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Shi, Junjie, Ma, Wei, Ying, Shi, Jiang, Lingxiao, liu, Yang, Du, Bo
Format:	Preprint
Publié:	2025
Sujets:	Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Project MPG: towards a generalized performance benchmark for LLM capabilities
par: Spangher, Lucas, et autres
Publié: (2024)

How predictable is language model benchmark performance?
par: Owen, David
Publié: (2024)

Sloth: scaling laws for LLM skills to predict multi-benchmark performance across families
par: Polo, Felipe Maia, et autres
Publié: (2024)

Agents Are All You Need for LLM Unlearning
par: Sanyal, Debdeep, et autres
Publié: (2025)

Forget Attention: Importance-Aware Attention Is All You Need
par: Shin, Soohyeong, et autres
Publié: (2026)

Fast Wrong-way Cycling Detection in CCTV Videos: Sparse Sampling is All You Need
par: Xu, Jing, et autres
Publié: (2024)

Rethinking Testing for LLM Applications: Characteristics, Challenges, and a Lightweight Interaction Protocol
par: Ma, Wei, et autres
Publié: (2025)

Agent psychometrics: Task-level performance prediction in agentic coding benchmarks
par: Ge, Chris, et autres
Publié: (2026)

What drives performance in molecular MPNNs? An operator-level factorial benchmark
par: Jiao, Panyu, et autres
Publié: (2026)

All You Need in Knowledge Distillation Is a Tailored Coordinate System
par: Zhou, Junjie, et autres
Publié: (2024)

PsychBench: A comprehensive and professional benchmark for evaluating the performance of LLM-assisted psychiatric clinical practice
par: Liu, Shuyu, et autres
Publié: (2025)

Dynamic benchmarking framework for LLM-based conversational data capture
par: Aluffi, Pietro Alessandro, et autres
Publié: (2025)

LLMzSzŁ: a comprehensive LLM benchmark for Polish
par: Jassem, Krzysztof, et autres
Publié: (2025)

Reasoning Is All You Need for Urban Planning AI
par: Yang, Sijie, et autres
Publié: (2025)

One Permutation Is All You Need: Fast, Reliable Variable Importance and Model Stress-Testing
par: Dorador, Albert
Publié: (2025)

Communication is All You Need: Persuasion Dataset Construction via Multi-LLM Communication
par: Ma, Weicheng, et autres
Publié: (2025)

Attention is All You Need Until You Need Retention
par: Yaslioglu, M. Murat
Publié: (2025)

Common Sense Is All You Need
par: Latapie, Hugo
Publié: (2025)

Training on the Benchmark Is Not All You Need
par: Ni, Shiwen, et autres
Publié: (2024)

Is Diversity All You Need for Scalable Robotic Manipulation?
par: Shi, Modi, et autres
Publié: (2025)

Easy Samples Are All You Need: Self-Evolving LLMs via Data-Efficient Reinforcement Learning
par: Yu, Zhiyin, et autres
Publié: (2026)

Contrast Is All You Need
par: Kilic, Burak, et autres
Publié: (2023)

Context is All You Need
par: Delanois, Jean Erik, et autres
Publié: (2026)

Polish-English medical knowledge transfer: A new benchmark and results
par: Grzybowski, Łukasz, et autres
Publié: (2024)

Information Gain Is Not All You Need
par: Ericson, Ludvig, et autres
Publié: (2025)

CXMArena: Unified Dataset to benchmark performance in realistic CXM Scenarios
par: Garg, Raghav, et autres
Publié: (2025)

Suvach -- Generated Hindi QA benchmark
par: Narayanan, Vaishak, et autres
Publié: (2024)

The role of positional encodings in the ARC benchmark
par: Costa, Guilherme H. Bandeira, et autres
Publié: (2025)

Memory augment is All You Need for image restoration
par: Zhang, Xiao Feng, et autres
Publié: (2023)

A benchmark multimodal oro-dental dataset for large vision-language models
par: Lv, Haoxin, et autres
Publié: (2025)

Exploitation Is All You Need... for Exploration
par: Rentschler, Micah, et autres
Publié: (2025)

[MASK] is All You Need
par: Hu, Vincent Tao, et autres
Publié: (2024)

Blending Is All You Need: Cheaper, Better Alternative to Trillion-Parameters LLM
par: Lu, Xiaoding, et autres
Publié: (2024)

Are Tools All We Need? Unveiling the Tool-Use Tax in LLM Agents
par: Zhang, Kaituo, et autres
Publié: (2026)

Belief Is All You Need: Modeling Narrative Archetypes in Conspiratorial Discourse
par: Shimgekar, Soorya Ram, et autres
Publié: (2025)

Pipeline Parallelism is All You Need for Optimized Early-Exit Based Self-Speculative Decoding
par: Li, Ruanjun, et autres
Publié: (2025)

Looped ReLU MLPs May Be All You Need as Practical Programmable Computers
par: Liang, Yingyu, et autres
Publié: (2024)

AstroCompress: A benchmark dataset for multi-purpose compression of astronomical data
par: Truong, Tuan, et autres
Publié: (2025)

Demystifying the Paradox of Importance Sampling with an Estimated History-Dependent Behavior Policy in Off-Policy Evaluation
par: Zhou, Hongyi, et autres
Publié: (2025)

Cooperation Is All You Need
par: Adeel, Ahsan, et autres
Publié: (2023)