Papers › On the State of the Art of Evaluation in Neural Language Models

On the State of the Art of Evaluation in Neural Language Models

18 Jul 2017ICLR 2018 1arXiv:1707.05589archive 2025-07-28

Gábor Melis, Chris Dyer, Phil Blunsom

Ongoing innovations in recurrent neural network architectures have provided a steady influx of apparently state-of-the-art results on language modelling benchmarks. However, these have been evaluated using differing code bases and limited computational resources, which represent uncontrolled sources of experimental variation. We reevaluate several popular architectures and regularisation methods with large-scale automatic black-box hyperparameter tuning and arrive at the somewhat surprising conclusion that standard LSTM architectures, when properly regularised, outperform more recent models. We establish a new state of the art on the Penn Treebank and Wikitext-2 corpora, as well as strong baselines on the Hutter Prize dataset.

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

deepmind/lamb mentioned on GitHubtfApache-2.0 report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Language Modelling

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Language Modelling WikiText-2 Melis et al. (2017) - 1-layer LSTM (tied) Number of params 24M #32 of 38 Archive leaderboard report
Language Modelling WikiText-2 Melis et al. (2017) - 1-layer LSTM (tied) Test perplexity 65.9 #32 of 38 Archive leaderboard report
Language Modelling WikiText-2 Melis et al. (2017) - 1-layer LSTM (tied) Validation perplexity 69.3 #32 of 38 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections