AI Paper Atlas

Research map / Benchmarks / PIQA

PIQA leaderboard

70 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on PIQA. Best accuracy: 91.8 (InfoSteer intervention, NeurIPS 2025).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

PIQA · accuracy

70 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
191.8InfoSteer intervention
Gemma-2-9B
Steering Information Utility in Key-Value Memory for Language Model Post-Training
NeurIPS 2025 · Table 1
Instruction and fine-tuning
290.6FILet
Llama3-8B
Learning in the Fisher Subspace: A Guided Initialization for LoRA Fine-Tuning
ICML 2026 · Table 1
Pruning and efficient tuning
390.29Dynamic CoT (transfer)
GPT-4o-mini and GPT-4o
Token Signature: Predicting Chain-of-Thought Gains with Token Decoding Feature in Large Language Models
ICML 2025 · Table: Cross-model transfer ef
Reasoning and chain of thought
489.07Paraphrase
GPT-3.5
Do as I do (Safely): Mitigating Task-Specific Fine-tuning Risks in Large Language Models
ICLR 2025 · Table: GPT-3.5 Fine-tuning on
Instruction and fine-tuning
587.5UniTE
OpenChat + Mistral + DeepSeek
Determine-Then-Ensemble: Necessity of Top-k Union for Large Language Model Ensembling
ICLR 2025 · Table: Results of ensembling o
Language, safety and interpretability
687.2KnowProxy
Llama-3.2-3B + Llama-3.2-1B proxy
KnowProxy: Adapting Large Language Models by Knowledge-guided Proxy
ICLR 2026 · Table 4
Instruction and fine-tuning
786LoRA-DA
LLaMA 2-7B
LoRA-DA: Data-Aware Initialization for Low-Rank Adaptation via Asymptotic Analysis
ICML 2026 · Table 1
Optimizers
885.9LEPA
Llama 3 8B Instruct
Learning to Plan Before Answering: Self-Teaching LLMs to Learn Abstract Plans for Problem Solving
ICLR 2025 · Table 1
Agents, code and math
985.85FreeCache + Qwen-7B-Instruct guided
Dream-7B-Instruct
FlashDLM: Accelerating Diffusion Language Model Inference via Efficient KV Caching and Guided Diffusion
ICLR 2026 · Table 2
Fast diffusion sampling
1085.3S-MeZO
Mistral-7B
Sparse MeZO: Less Parameters for Better Performance in Zeroth-Order LLM Fine-Tuning
NeurIPS 2025 · Table: Accuracy of Fine-Tuning
Pruning and efficient tuning
60 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.