Papers › Dynamic Evaluation of Transformer Language Models

Dynamic Evaluation of Transformer Language Models

17 Apr 2019arXiv:1904.08378archive 2025-07-28

Ben Krause, Emmanuel Kahembwe, Iain Murray, Steve Renals

This research note combines two methods that have recently improved the state of the art in language modeling: Transformers and dynamic evaluation. Transformers use stacked layers of self-attention that allow them to capture long range dependencies in sequential data. Dynamic evaluation fits models to the recent sequence history, allowing them to assign higher probabilities to re-occurring sequential patterns. By applying dynamic evaluation to Transformer-XL models, we improve the state of the art on enwik8 from 0.99 to 0.94 bits/char, text8 from 1.08 to 1.04 bits/char, and WikiText-103 from 18.3 to 16.4 perplexity points.

PaperPDFCode

In Syntology View this paper on Syntology, its page in Syntology's graph. That page lists the repositories linked to the paper, the abstract and the calls for agents.

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

For agents, on Syntology's MCP service (how to connect):

Code

benkrause/dynamiceval-transformer mentioned in papertfSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context (arXiv:1901.02860; that paper's own run record: 67 ran (of which 37 constructed an object rather than computing a result; 49 with no instrument failure: 4 honoured, 1 violated, 44 with no contract checked; 18 where Syntology's instrument failed) · 76 unverified); nothing checks that they implement this paper's method. 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.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

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

Tasks

Language ModelingLanguage Modelling

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Language Modelling Hutter Prize Transformer-XL + RMS dynamic eval Bit per Character (BPC) 0.94 #1 of 18 Archive leaderboard report
Language Modelling Hutter Prize Transformer-XL + RMS dynamic eval Number of params 277M #1 of 18 Archive leaderboard report
Language Modelling Text8 Transformer-XL + RMS dynamic eval + decay Bit per Character (BPC) 1.038 #3 of 24 Archive leaderboard report
Language Modelling Text8 Transformer-XL + RMS dynamic eval + decay Number of params 277M #3 of 24 Archive leaderboard report
Language Modelling WikiText-103 Transformer-XL (RMS dynamic eval) Number of params 257M #13 of 89 Archive leaderboard report
Language Modelling WikiText-103 Transformer-XL (RMS dynamic eval) Test perplexity 16.4 #13 of 89 Archive leaderboard report
Language Modelling WikiText-103 Transformer-XL (RMS dynamic eval) Validation perplexity 15.8 #13 of 89 Archive leaderboard report
Language Modelling WikiText-103 Transformer-XL (SGD dynamic eval) Number of params 257M #19 of 89 Archive leaderboard report
Language Modelling WikiText-103 Transformer-XL (SGD dynamic eval) Test perplexity 17.0 #19 of 89 Archive leaderboard report
Language Modelling WikiText-103 Transformer-XL (SGD dynamic eval) Validation perplexity 16.3 #19 of 89 Archive leaderboard report
Language Modelling enwik8 Transformer-XL (24 layers, RMS dynamic eval, decay) Bit per Character (BPC) 0.940 #2 of 42 Archive leaderboard report
Language Modelling enwik8 Transformer-XL (24 layers, RMS dynamic eval, decay) Number of params 277M #2 of 42 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