Skip to content
AI News HubLIVE
Original source2 min read

Richness Is Not Evenness: Ordered Renyi Crossings under Self-Training

Summary

This paper studies recursive self-training of language models and finds that as training temperature rises, different Renyi entropies (min-entropy, collision entropy, Shannon entropy, and observed richness) cross from loss to gain at distinct temperatures in a consistent order, which persists across various settings.

SourcearXiv Computational LinguisticsAuthor: Pranjal Srivastava
Richness Is Not Evenness: Ordered Renyi Crossings under Self-Training
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 6 Nov 2026]

Title:Richness Is Not Evenness: Ordered Renyi Crossings under Self-Training

View a PDF of the paper titled Richness Is Not Evenness: Ordered Renyi Crossings under Self-Training, by Pranjal Srivastava

View PDF HTML (experimental)

Abstract:Recursive self-training of language models, a setting studied in model collapse, can reduce vocabulary while making frequent words, mostly function words, less dominant. We fine-tune English GPT-2 on its own output for nine generations, vary the training-data sampling temperature, and measure every model with one fixed sampler. As training temperature rises, Renyi entropy changes turn from loss to gain at distinct temperatures: min-entropy first, then collision entropy, Shannon entropy and observed richness. Between the outer crossings, word types fall while the most-common word becomes less dominant. Untruncated training at T = 1 lies in this interval: at 13,000 words, generation 9 has 24% fewer types than generation 0 yet collision entropy is 0.91 bits higher. The ordering persists with a finer grid, different measurement samplers, a fixed word budget and five fresh training seeds. Min-entropy again crosses before richness in four further settings (in Hindi only at interpolated roots) and in 12 of 13 distinct measurements under eight text representations. Next-token identities and a categorical simulation isolate decoding and sampling effects; the crossing order remains empirical. Fine-tuning on human text can produce the same split while likelihood improves, and no single order's sign reliably identifies worse likelihood across the tested settings. We recommend reporting entropy profiles at stated budgets with an independent quality measure.

Subjects:

Computation and Language (cs.CL)

Cite as: arXiv:2611.08892 [cs.CL]

(or arXiv:2611.08892v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2611.08892

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Pranjal Srivastava [view email] [v1] Fri, 6 Nov 2026 02:31:07 UTC (217 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Richness Is Not Evenness: Ordered Renyi Crossings under Self-Training, by Pranjal Srivastava

View PDF

HTML (experimental)

TeX Source

view license

Additional Features

Audio Summary

Current browse context:

cs.CL

new | recent | 2026-11

Change to browse by:

cs

References & Citations

SciX

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • During self-training, as training temperature increases, min-entropy, collision entropy, Shannon entropy, and observed richness transition from loss to gain in a fixed order.
  • At T=1 with untruncated training, generation 9 has 24% fewer word types than generation 0 but collision entropy is 0.91 bits higher, showing richness can fall while evenness improves.
  • The crossing order holds with a finer grid, different measurement samplers, a fixed word budget, and five fresh training seeds, and is observed in multiple languages and representations.
  • Fine-tuning on human text can produce the same split while likelihood improves, so no single entropy order's sign reliably indicates worse likelihood.

Highlights and analysis are generated automatically and may contain errors. Check the original source.