# Anticipatory Evaluation of Language Models: Predicting LLM Benchmark Scores from Descriptions

arXiv: https://arxiv.org/abs/2509.20645 · HTML: https://mlsanity.com/paper/2509.20645/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 54 of its 67 references that are in the map.
- First posted on arXiv 25 Sep 2025.

## Builds on (most cited first, 30 of 54)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [XGBoost: A Scalable Tree Boosting System](https://mlsanity.com/paper/1603.02754/) (2016)
- [GPT-4o System Card](https://mlsanity.com/paper/2410.21276/) (2024)
- [Training Compute-Optimal Large Language Models](https://mlsanity.com/paper/2203.15556/) (2022)
- [Qwen2 Technical Report](https://mlsanity.com/paper/2407.10671/) (2024)
- [MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark](https://mlsanity.com/paper/2406.01574/) (2024)
- [Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models](https://mlsanity.com/paper/2506.05176/) (2025)
- [Chronos: Learning the Language of Time Series](https://mlsanity.com/paper/2403.07815/) (2024)
- [Time-LLM: Time Series Forecasting by Reprogramming Large Language Models](https://mlsanity.com/paper/2310.01728/) (2023)
- [A decoder-only foundation model for time-series forecasting](https://mlsanity.com/paper/2310.10688/) (2023)
- [Large Language Models Are Zero-Shot Time Series Forecasters](https://mlsanity.com/paper/2310.07820/) (2023)
- [tinyBenchmarks: evaluating LLMs with fewer examples](https://mlsanity.com/paper/2402.14992/) (2024)
- [Improving Text Embeddings with Large Language Models](https://mlsanity.com/paper/2401.00368/) (2023)
- [Task2Vec: Task Embedding for Meta-Learning](https://mlsanity.com/paper/1902.03545/) (2019)
- [TEMPO: Prompt-based Generative Pre-trained Transformer for Time Series Forecasting](https://mlsanity.com/paper/2310.04948/) (2023)
- [Are Language Models Actually Useful for Time Series Forecasting?](https://mlsanity.com/paper/2406.16964/) (2024)
- [ForecastBench: A Dynamic Benchmark of AI Forecasting Capabilities](https://mlsanity.com/paper/2409.19839/) (2024)
- [Approaching Human-Level Forecasting with Language Models](https://mlsanity.com/paper/2402.18563/) (2024)
- [The ConceptARC Benchmark: Evaluating Understanding and Generalization in the ARC Domain](https://mlsanity.com/paper/2305.07141/) (2023)
- [Reasoning Models Better Express Their Confidence](https://mlsanity.com/paper/2505.14489/) (2025)
- [Forecasting Future World Events with Neural Networks](https://mlsanity.com/paper/2206.15474/) (2022)
- [From Words to Numbers: Your Large Language Model Is Secretly A Capable Regressor When Given In-Context Examples](https://mlsanity.com/paper/2404.07544/) (2024)
- [Comparing Humans, GPT-4, and GPT-4V On Abstraction and Reasoning Tasks](https://mlsanity.com/paper/2311.09247/) (2023)
- [LLM Processes: Numerical Predictive Distributions Conditioned on Natural Language](https://mlsanity.com/paper/2405.12856/) (2024)
- [Identification of Tasks, Datasets, Evaluation Metrics, and Numeric Scores for Scientific Leaderboards Construction](https://mlsanity.com/paper/1906.09317/) (2019)
- [Predicting Performance for Natural Language Processing Tasks](https://mlsanity.com/paper/2005.00870/) (2020)
- [AxCell: Automatic Extraction of Results from Machine Learning Papers](https://mlsanity.com/paper/2004.14356/) (2020)

## Built on by (most cited first, 2 of 2)

- [Reinforcement Learning over Predictive Distributions for LLM Regression](https://mlsanity.com/paper/2605.20740/) (2026)
- [Balance of Benchmarks: Semantic Density Reweighting for Task-Conditioned Model Comparison](https://mlsanity.com/paper/2608.30044/) (2026)

## Abstract (arXiv)

Modern ML research increasingly faces an evaluation bottleneck: building benchmarks and running experiments is costly, slowing iteration. We ask how much can be inferred about evaluation outcomes before experiments, given only a natural-language description of the task and setup. We call this setting anticipatory evaluation. We introduce EVALCAST, a literature-grounded resource of 2,290 description-performance pairs from 1,519 arXiv papers, including a test set of papers published after the model's knowledge cutoff. We evaluate two tasks: predicting absolute performance through regression and comparing benchmark difficulty through pairwise prediction. Experiments with frontier LLMs show that accurate regression remains challenging, while pairwise predictions can recover relative performance ordering under matched settings. Comparable performance on post-cutoff and streaming evaluations suggests that the observed signal is not solely attributable to previously reported results. Overall, EVALCAST provides a resource and evaluation framework for studying anticipatory evaluation and its potential use in prioritizing benchmarks.