pith. sign in

arxiv: 2606.12422 · v1 · pith:R6ZCCN4Onew · submitted 2026-05-08 · 💻 cs.CY · cs.AI· cs.HC

Creating and Evaluating K-12 GenAI Assessment Graders Through Context Engineering

Zewei Tian , Alex Liu , Lief Esbenshade , Michael Xiao , Zachary Zhang , Yulia L\'apicus , Thomas Han , Kevin He
show 1 more author
Min Sun
This is my paper

Pith reviewed 2026-06-30 22:51 UTC · model grok-4.3

classification 💻 cs.CY cs.AIcs.HC
keywords LLM gradersK-12 assessmentautomated scoringcontext engineeringinterrater agreementMCASgenerative AI in educationstandards-based grading
0
0 comments X

The pith

Context-engineered LLMs achieve substantial agreement with human raters on K-12 math and science assessments.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper tests whether off-the-shelf large language models can grade K-12 student responses reliably when given context and prompt engineering. It measures agreement with human raters on actual Massachusetts Comprehensive Assessment System items in mathematics, science, and English language arts. Larger models reach substantial alignment in math and science but show more variable results in ELA. Teacher and student feedback favors the narrative comments over the numerical scores. The work therefore positions these graders as useful components in hybrid systems rather than replacements for teacher judgment.

Core claim

By applying context and prompt engineering to commercially available foundation models, LLM-based graders can produce scores that align substantially with human raters on mathematics and science items from the MCAS, as measured by Quadratic Weighted Kappa and PRMSE, though results vary in ELA; the approach works without fine-tuning and supports hybrid human-AI assessment models.

What carries the argument

The context-engineered LLM grader that scores student responses against rubrics using foundation models such as Claude Sonnet 4 and GPT-5.

If this is right

  • Larger-parameter models produce higher agreement with humans in mathematics and science.
  • Agreement is lower and more variable in English language arts.
  • Narrative feedback from the models receives stronger acceptance than numerical scores.
  • Hybrid systems that combine LLM output with teacher review can reduce workload while preserving professional judgment.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same context-engineering approach could be tested on other state assessments or classroom rubrics to check consistency.
  • Over time, teacher time might shift from initial scoring to reviewing and interpreting AI-generated comments.
  • Equity effects across different student populations remain untested in the reported data.

Load-bearing premise

That carefully designed prompts and context on generic foundation models produce consistent, reliable scores without subject-specific training data or fine-tuning.

What would settle it

A new study using different student responses or rubrics that reports low Quadratic Weighted Kappa values between the same LLM setup and human raters.

Figures

Figures reproduced from arXiv: 2606.12422 by Alex Liu, Kevin He, Lief Esbenshade, Michael Xiao, Min Sun, Thomas Han, Yulia L\'apicus, Zachary Zhang, Zewei Tian.

Figure 1
Figure 1. Figure 1: Illustration of Context Engineering process for LLM Grading [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
read the original abstract

The integration of large language models (LLMs) into educational assessment represents a transformative shift in classroom grading practices. While automated scoring systems and machine learning techniques have existed for decades, generative AI (GenAI) now enables educators to implement standards-based grading (SBG) with unprecedented efficiency and scale. This paper examines the theoretical foundations and evaluates an LLM grader that uses commercially available foundation models with context and prompt engineering to score student work against a rubric. Drawing on an empirical interrater agreement study using Massachusetts Comprehensive Assessment System (MCAS) data, we observed the Quadratic Weighted Kappa (QWK) and Proportional Reduction in Mean-Squared Error (PRMSE) across mathematics, science, and ELA, using Claude Sonnet 4, Haiku 4.5, GPT-5, and GPT-5 Mini. The results demonstrate that LLM graders, especially when based on foundational models with more parameters, achieve substantial agreement with human raters in mathematics and science assessments, while the performances vary in ELA, suggesting generic foundation models can be effective at scoring in given contexts. Additional analysis of teacher and student feedback reveals strong acceptance of AI-generated narrative feedback but skepticism toward numerical scores, suggesting that LLMs function most effectively as formative tools rather than summative evaluators. Our findings indicate that thoughtfully designed hybrid models that combine AI efficiency with teacher judgment can reduce workload, enhance feedback quality, and support equitable assessment practices without displacing professional expertise.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript claims that context- and prompt-engineered LLM graders built on commercial foundation models (Claude Sonnet 4, Haiku 4.5, GPT-5, GPT-5 Mini) achieve substantial agreement with human raters on MCAS K-12 items in mathematics and science, as measured by QWK and PRMSE, with more variable results in ELA. It further reports that teachers and students accept AI-generated narrative feedback but remain skeptical of numerical scores, concluding that such systems are best used as formative tools within hybrid human-AI assessment workflows.

Significance. If the reported agreement metrics prove robust once anchored to human baselines and supplied with full methodological detail, the work would offer a practical demonstration that generic foundation models can support standards-based grading at scale without subject-specific fine-tuning. This could inform workload-reduction strategies in K-12 settings and contribute to discussions of equitable, hybrid assessment. The absence of a human-human reference and of basic statistical reporting currently limits the strength of that contribution.

major comments (2)
  1. [Abstract / Methodology] Abstract and methodology description: the claim that LLM graders achieve 'substantial agreement' with human raters cannot be evaluated without a human-human interrater QWK baseline computed on the identical MCAS items and rubrics. Without this ceiling, reported QWK values remain unanchored and the interpretation of 'substantial' is unsupported.
  2. [Results] Results section: the abstract states QWK and PRMSE values across subjects and models but supplies no sample sizes, exact prompts, data filtering rules, confidence intervals, or statistical tests. These omissions prevent assessment of the precision and replicability of the central empirical claims.
minor comments (1)
  1. [Abstract] The abstract would benefit from a brief statement of the number of student responses or items scored to convey study scale.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for these constructive comments, which identify key areas where additional context and transparency can strengthen the manuscript. We address each major comment below.

read point-by-point responses

  1. Referee: [Abstract / Methodology] Abstract and methodology description: the claim that LLM graders achieve 'substantial agreement' with human raters cannot be evaluated without a human-human interrater QWK baseline computed on the identical MCAS items and rubrics. Without this ceiling, reported QWK values remain unanchored and the interpretation of 'substantial' is unsupported.

    Authors: We agree that a human-human baseline would aid interpretation of the reported QWK values. However, the MCAS released items in our study follow standard operational scoring procedures that provide only a single human rater per response; we do not have access to multiple independent human ratings on the identical items and therefore cannot compute such a baseline. We will revise the manuscript to explicitly acknowledge this limitation, clarify that our QWK values measure agreement with the operational human scores rather than an empirical ceiling, and anchor the term 'substantial' to conventional benchmarks in the educational measurement literature (e.g., Landis & Koch). revision: partial

  2. Referee: [Results] Results section: the abstract states QWK and PRMSE values across subjects and models but supplies no sample sizes, exact prompts, data filtering rules, confidence intervals, or statistical tests. These omissions prevent assessment of the precision and replicability of the central empirical claims.

    Authors: We appreciate this observation. While the full methods and results sections contain sample sizes and basic methodological information, we acknowledge that the abstract and certain results subsections omit exact prompts, data filtering rules, confidence intervals, and statistical tests. We will revise to add these details: include sample sizes in the abstract or prominently in results, move the exact prompts to a new appendix, describe filtering rules explicitly, and report confidence intervals along with any applicable statistical tests. revision: yes

standing simulated objections not resolved
  • Human-human interrater QWK baseline on the identical MCAS items (unavailable due to single-rater operational scoring)

Circularity Check

0 steps flagged

Empirical evaluation study with no derivation chain or self-referential elements

full rationale

The paper is a direct empirical comparison of LLM graders (via context engineering on foundation models) against human raters on MCAS items, reporting QWK and PRMSE values across subjects. No equations, parameter fitting, or predictive derivations are described; the central claim rests on observed agreement metrics rather than any reduction to self-defined inputs or self-citation chains. The study is self-contained against the external human rating benchmark and does not invoke uniqueness theorems, ansatzes, or renamings that collapse back to the paper's own choices.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no equations, fitted constants, or postulated entities; ledger left empty.

pith-pipeline@v0.9.1-grok · 5817 in / 1031 out tokens · 24933 ms · 2026-06-30T22:51:30.737552+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

272 extracted references · 139 canonical work pages · 9 internal anchors

  1. [1]

    , year =

    Scheurman, Geoffrey and Newmann, Fred M. , year =. Authentic intellectual work in social studies:. Social Education , publisher =

  2. [2]

    and Bryk, Anthony S

    Newmann, Fred M. and Bryk, Anthony S. and Nagaoka, Jenny K. , year =. Authentic

  3. [3]

    Nature Machine Intelligence , author =

    Principles alone cannot guarantee ethical. Nature Machine Intelligence , author =. 2019 , keywords =. doi:10.1038/s42256-019-0114-4 , abstract =

    work page doi:10.1038/s42256-019-0114-4 2019

  4. [5]

    International Journal of Medical Informatics , author =

    Finding warning markers:. International Journal of Medical Informatics , author =. 2020 , keywords =. doi:10.1016/j.ijmedinf.2020.104137 , abstract =

    work page doi:10.1016/j.ijmedinf.2020.104137 2020

  5. [6]

    , month = dec, year =

    Gillies, Robyn M. , month = dec, year =. Enhancing

  6. [7]

    Smart Learning Environments , author =

    Personalized adaptive learning: an emerging pedagogical approach enabled by a smart learning environment , volume =. Smart Learning Environments , author =. 2019 , keywords =. doi:10.1186/s40561-019-0089-y , abstract =

    work page doi:10.1186/s40561-019-0089-y 2019

  7. [8]

    Harnessing the

    Alam, Ashraf , editor =. Harnessing the. Intelligent. 2023 , keywords =. doi:10.1007/978-981-99-1767-9_42 , abstract =

    work page doi:10.1007/978-981-99-1767-9_42 2023

  8. [9]

    Journal of Educational Psychology , author =

    Using adaptive learning technologies to personalize instruction to student interests:. Journal of Educational Psychology , author =. 2013 , keywords =. doi:10.1037/a0031882 , abstract =

    work page doi:10.1037/a0031882 2013

  9. [10]

    IEEE Signal Processing Magazine , author =

    Personalized. IEEE Signal Processing Magazine , author =. 2021 , pages =. doi:10.1109/MSP.2021.3055032 , abstract =

    work page doi:10.1109/msp.2021.3055032 2021

  10. [12]

    Makinae, Naomichi , editor =. The. Theory and. 2019 , keywords =. doi:10.1007/978-3-030-04031-4_9 , abstract =

    work page doi:10.1007/978-3-030-04031-4_9 2019

  11. [13]

    Education and Information Technologies , author =

    Large language models in education:. Education and Information Technologies , author =. 2023 , keywords =. doi:10.1007/s10639-023-11834-1 , abstract =

    work page doi:10.1007/s10639-023-11834-1 2023

  12. [14]

    Studies in Educational Evaluation , author =

    Classroom observation for evaluating and improving teaching:. Studies in Educational Evaluation , author =. 2016 , pages =

    2016

  13. [17]

    Innovations in Education and Teaching International , author =

    Overlooking the conceptual framework , volume =. Innovations in Education and Teaching International , author =. 2007 , pages =. doi:10.1080/14703290601081407 , language =

    work page doi:10.1080/14703290601081407 2007

  14. [18]

    The School Review , author =

    The. The School Review , author =. doi:10.1086/442404 , language =

    work page doi:10.1086/442404

  15. [19]

    Red teaming chatgpt via jailbreaking:

    Zhuo, Terry Yue and Huang, Yujin and Chen, Chunyang and Xing, Zhenchang , year =. Red teaming chatgpt via jailbreaking:. arXiv preprint arXiv:2301.12867 , publisher =

    arXiv

  16. [20]

    and Rodríguez, Roberto J

    Cardona, Miguel A. and Rodríguez, Roberto J. and Ishmael, Kristina , year =. Artificial

  17. [21]

    Language and Education , author =

    An analysis of the forms of teacher-student dialogue that are most productive for learning , volume =. Language and Education , author =. 2023 , pages =. doi:10.1080/09500782.2021.1956943 , language =

    work page doi:10.1080/09500782.2021.1956943 2023

  18. [22]

    Retrieved from Glossary of Artificial Intelligence Terms for Educators–CIRCLS , author =

    Glossary of. Retrieved from Glossary of Artificial Intelligence Terms for Educators–CIRCLS , author =

  19. [23]

    Annual Review of Information Science and Technology (ARIST) , author =

    Natural. Annual Review of Information Science and Technology (ARIST) , author =. 2003 , note =

    2003

  20. [24]

    Technology, Knowledge and Learning , author =

    Educational. Technology, Knowledge and Learning , author =. 2014 , file =. doi:10.1007/s10758-014-9223-7 , abstract =

    work page doi:10.1007/s10758-014-9223-7 2014

  21. [25]

    and Elmore, Richard F

    City, Elizabeth A. and Elmore, Richard F. and Fiarman, Sarah E. and Teitel, Lee , year =. Instructional rounds in education , volume =

  22. [26]

    Draft manuscript , author =

    Improving the instructional core , url =. Draft manuscript , author =. 2008 , file =

    2008

  23. [27]

    , month = dec, year =

    Scribner, Jay Paredes and Donaldson, Joe F. , month = dec, year =. The. Educational Administration Quarterly , publisher =. doi:10.1177/00131610121969442 , abstract =

    work page doi:10.1177/00131610121969442

  24. [28]

    Learning

    Abrahamson, Dor and Sánchez-García, Raúl , month = apr, year =. Learning. Journal of the Learning Sciences , publisher =. doi:10.1080/10508406.2016.1143370 , abstract =

    work page doi:10.1080/10508406.2016.1143370 2016

  25. [29]

    Educational Technology Research and Development , author =

    Technology enhanced feedback tools as a knowledge management mechanism for supporting professional growth and school reform , volume =. Educational Technology Research and Development , author =. 2011 , pages =. doi:10.1007/S11423-011-9201-X , abstract =

    work page doi:10.1007/s11423-011-9201-x 2011

  26. [30]

    and Stoolmiller, Mike and Kennedy, Patrick C

    Doabler, Christian T. and Stoolmiller, Mike and Kennedy, Patrick C. and Nelson, Nancy J. and Clarke, Ben and Gearin, Brian and Fien, Hank and Smolkowski, Keith and Baker, Scott K. , month = jun, year =. Do. Assessment for Effective Intervention , publisher =. doi:10.1177/1534508418758364 , abstract =

    work page doi:10.1177/1534508418758364

  27. [31]

    Journal of Research in Science Teaching , author =

    Impact of project-based curriculum materials on student learning in science:. Journal of Research in Science Teaching , author =. 2015 , note =. doi:10.1002/tea.21263 , abstract =

    work page doi:10.1002/tea.21263 2015

  28. [32]

    Neuro‐linguistic programming as an innovation in education and teaching , volume =

    Tosey, Paul and Mathison, Jane , month = aug, year =. Neuro‐linguistic programming as an innovation in education and teaching , volume =. Innovations in Education and Teaching International , publisher =. doi:10.1080/14703297.2010.498183 , abstract =

    work page doi:10.1080/14703297.2010.498183 2010

  29. [33]

    Studies in Philosophy and Education , author =

    Deliberative. Studies in Philosophy and Education , author =. 2008 , keywords =. doi:10.1007/s11217-007-9071-1 , abstract =

    work page doi:10.1007/s11217-007-9071-1 2008

  30. [34]

    Measuring

    Liu, Jing and Cohen, Julie , month = dec, year =. Measuring. Educational Evaluation and Policy Analysis , publisher =. doi:10.3102/01623737211009267 , abstract =

    work page doi:10.3102/01623737211009267

  31. [35]

    Danielson, Charlotte , year =

  32. [36]

    The framework for teaching evaluation instrument

    Danielson, Charlotte , year =. The framework for teaching evaluation instrument. , url =

  33. [37]

    Yang, Xi and Zhang, Lishan and Yu, Shengquan , editor =. Can. Artificial. 2017 , note =. doi:10.1007/978-3-319-61425-0_72 , urldate =

    work page doi:10.1007/978-3-319-61425-0_72 2017

  34. [38]

    and Sutinen, E

    Kakkonen, T. and Sutinen, E. , year =. Automatic assessment of the content of essays based on course materials , isbn =. doi:10.1109/ITRE.2004.1393660 , abstract =

    work page doi:10.1109/itre.2004.1393660 2004

  35. [39]

    Intelligent

    Wang, Zining and Liu, Jianli and Dong, Ruihai , month = nov, year =. Intelligent. 2018 5th. doi:10.1109/CCIS.2018.8691244 , abstract =

    work page doi:10.1109/ccis.2018.8691244 2018

  36. [40]

    Leading the instructional core , journal =

    Elmore, Richard , year =. Leading the instructional core , journal =

  37. [41]

    Key issues of lesson study in

    Saito, Eisuke , month = nov, year =. Key issues of lesson study in. Professional Development in Education , publisher =. doi:10.1080/19415257.2012.668857 , abstract =

    work page doi:10.1080/19415257.2012.668857 2012

  38. [42]

    Promoting rich discussions in mathematics classrooms:

    Jacobs, Jennifer and Scornavacco, Karla and Harty, Charis and Suresh, Abhijit and Lai, Vivian and Sumner, Tamara , year =. Promoting rich discussions in mathematics classrooms:. Teaching and Teacher Education , publisher =

  39. [43]

    Suresh, Abhijit and Jacobs, Jennifer and Harty, Charis and Perkoff, Margaret and Martin, James H and Sumner, Tamara , file =. The

  40. [44]

    and Sumner, Tamara , month = apr, year =

    Suresh, Abhijit and Jacobs, Jennifer and Harty, Charis and Perkoff, Margaret and Martin, James H. and Sumner, Tamara , month = apr, year =. The. doi:10.48550/arXiv.2204.09652 , abstract =

    work page doi:10.48550/arxiv.2204.09652

  41. [45]

    Studies in Educational Evaluation , author =

    Classroom observation for evaluating and improving teaching:. Studies in Educational Evaluation , author =. 2016 , keywords =. doi:10.1016/j.stueduc.2016.03.002 , abstract =

    work page doi:10.1016/j.stueduc.2016.03.002 2016

  42. [46]

    Education Policy Analysis Archives , author =

    Using global observation protocols to inform research on teaching effectiveness and school improvement:. Education Policy Analysis Archives , author =. 2020 , keywords =. doi:10.14507/epaa.28.5012 , abstract =

    work page doi:10.14507/epaa.28.5012 2020

  43. [47]

    Fine-tuning

    Latif, Ehsan and Zhai, Xiaoming , month = dec, year =. Fine-tuning

  44. [48]

    and Demszky, Dorottya , month = jun, year =

    Wang, Rose E. and Demszky, Dorottya , month = jun, year =. Is. doi:10.48550/arXiv.2306.03090 , abstract =

    work page doi:10.48550/arxiv.2306.03090

  45. [49]

    and Stone, Cathlyn and Kelly, Sean and Godley, Amanda and D'Mello, Sidney K

    Jensen, Emily and Dale, Meghan and Donnelly, Patrick J. and Stone, Cathlyn and Kelly, Sean and Godley, Amanda and D'Mello, Sidney K. , month = apr, year =. Toward. Proceedings of the 2020. doi:10.1145/3313831.3376418 , abstract =

    work page doi:10.1145/3313831.3376418 2020

  46. [50]

    and Jurafsky, Dan and Piech, Chris , month = may, year =

    Demszky, Dorottya and Liu, Jing and Hill, Heather C. and Jurafsky, Dan and Piech, Chris , month = may, year =. Can. Educational Evaluation and Policy Analysis , publisher =. doi:10.3102/01623737231169270 , abstract =

    work page doi:10.3102/01623737231169270

  47. [51]

    Innovative language teaching and learning at university: enhancing participation and collaboration , isbn =

    Goria, Cecilia and Speicher, Oranna and Stollhans, Sascha , month = jan, year =. Innovative language teaching and learning at university: enhancing participation and collaboration , isbn =

  48. [52]

    2016 , pages =

    Opening-up classroom discourse to promote and enhance active, collaborative and cognitively-engaging student learning experiences , isbn =. 2016 , pages =. doi:10.14705/rpnet.2016.000400 , author =

    work page doi:10.14705/rpnet.2016.000400 2016

  49. [53]

    Exploring

    Mercer, Neil and Hodgkinson, Steve , month = sep, year =. Exploring

  50. [54]

    Exploring talk in school , author =

    Culture, dialogue and learning:. Exploring talk in school , author =. 2008 , pages =

    2008

  51. [55]

    and Borsheim-Black, Carlin and Caughlan, Samantha and Heintz, Anne , year =

    Juzwik, Mary M. and Borsheim-Black, Carlin and Caughlan, Samantha and Heintz, Anne , year =. Inspiring dialogue:

  52. [56]

    International Journal of Educational Research , author =

    What the discourse tells us:. International Journal of Educational Research , author =. 2008 , keywords =. doi:10.1016/j.ijer.2009.01.001 , abstract =

    work page doi:10.1016/j.ijer.2009.01.001 2008

  53. [57]

    Manning, Christopher and Surdeanu, Mihai and Bauer, John and Finkel, Jenny and Bethard, Steven and McClosky, David , editor =. The. Proceedings of 52nd. 2014 , pages =. doi:10.3115/v1/P14-5010 , urldate =

    work page doi:10.3115/v1/p14-5010 2014

  54. [58]

    Surdeanu, Mihai and Hicks, Tom and Valenzuela-Escárcega, Marco Antonio , editor =. Two. Proceedings of the 2015. 2015 , pages =. doi:10.3115/v1/N15-3001 , urldate =

    work page doi:10.3115/v1/n15-3001 2015

  55. [59]

    Computers & Education , author =

    Automatic classification of activities in classroom discourse , volume =. Computers & Education , author =. 2014 , keywords =. doi:10.1016/j.compedu.2014.05.010 , abstract =

    work page doi:10.1016/j.compedu.2014.05.010 2014

  56. [60]

    and Bryk, Anthony S

    Newmann, Fred M. and Bryk, Anthony S. and Nagaoka, Jenny K. , month = jan, year =. Authentic

  57. [61]

    and Donnelly, Patrick and Nystrand, Martin and D’Mello, Sidney K

    Kelly, Sean and Olney, Andrew M. and Donnelly, Patrick and Nystrand, Martin and D’Mello, Sidney K. , month = oct, year =. Automatically. Educational Researcher , publisher =. doi:10.3102/0013189X18785613 , abstract =

    work page doi:10.3102/0013189x18785613

  58. [62]

    Demszky, Dorottya and Hill, Heather , month = may, year =. The. doi:10.48550/arXiv.2211.11772 , abstract =

    work page doi:10.48550/arxiv.2211.11772

  59. [63]

    Empowering

    Ruiz-Rojas, Lena Ivannova and Acosta-Vargas, Patricia and De-Moreta-Llovet, Javier and Gonzalez-Rodriguez, Mario , month = jan, year =. Empowering. Sustainability , publisher =. doi:10.3390/su151511524 , abstract =

    work page doi:10.3390/su151511524

  60. [64]

    Predicting

    Chen, Huanyi , month = jun, year =. Predicting

  61. [65]

    and Caballero, Marcos D

    Wilson, Joseph and Pollard, Benjamin and Aiken, John M. and Caballero, Marcos D. and Lewandowski, H. J. , month = jun, year =. Classification of open-ended responses to a research-based assessment using natural language processing , volume =. Physical Review Physics Education Research , publisher =. doi:10.1103/PhysRevPhysEducRes.18.010141 , abstract =

    work page doi:10.1103/physrevphyseducres.18.010141

  62. [66]

    World Journal of Surgery , author =

    Leveraging. World Journal of Surgery , author =. 2023 , pages =. doi:10.1007/s00268-023-07167-2 , language =

    work page doi:10.1007/s00268-023-07167-2 2023

  63. [67]

    Artificial Intelligence Review , author =

    Intelligent tutoring systems: an overview , volume =. Artificial Intelligence Review , author =. 1990 , keywords =. doi:10.1007/BF00168958 , abstract =

    work page doi:10.1007/bf00168958 1990

  64. [68]

    and Koedinger, Kenneth R

    Corbett, Albert T. and Koedinger, Kenneth R. and Anderson, John R. , editor =. Chapter 37 -. Handbook of. 1997 , pages =. doi:10.1016/B978-044481862-1.50103-5 , abstract =

    work page doi:10.1016/b978-044481862-1.50103-5 1997

  65. [69]

    Niakan Kalhori, Sharareh and Rakhshan, Mahnaz and Keikha, Leila and Ghazi Saeedi, Marjan , month = jan, year =

    Mousavinasab, Elham and Zarifsanaiey, Nahid and R. Niakan Kalhori, Sharareh and Rakhshan, Mahnaz and Keikha, Leila and Ghazi Saeedi, Marjan , month = jan, year =. Intelligent tutoring systems: a systematic review of characteristics, applications, and evaluation methods , volume =. Interactive Learning Environments , publisher =. doi:10.1080/10494820.2018....

    work page doi:10.1080/10494820.2018.1558257 2018

  66. [71]

    doi:10.48550/arXiv.2309.03241 , abstract =

    Yang, Zhen and Ding, Ming and Lv, Qingsong and Jiang, Zhihuan and He, Zehai and Guo, Yuyi and Bai, Jinfeng and Tang, Jie , month = sep, year =. doi:10.48550/arXiv.2309.03241 , abstract =

    work page doi:10.48550/arxiv.2309.03241

  67. [72]

    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

    Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina , month = may, year =. doi:10.48550/arXiv.1810.04805 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1810.04805

  68. [73]

    Wang et al

    Wang, Yidong and Yu, Zhuohao and Zeng, Zhengran and Yang, Linyi and Wang, Cunxiang and Chen, Hao and Jiang, Chaoya and Xie, Rui and Wang, Jindong and Xie, Xing and Ye, Wei and Zhang, Shikun and Zhang, Yue , month = jun, year =. doi:10.48550/arXiv.2306.05087 , abstract =

    work page doi:10.48550/arxiv.2306.05087

  69. [74]

    Improving

    Cochran, Keith and Cohn, Clayton and Rouet, Jean Francois and Hastings, Peter , editor =. Improving. Artificial. 2023 , keywords =. doi:10.1007/978-3-031-36272-9_18 , abstract =

    work page doi:10.1007/978-3-031-36272-9_18 2023

  70. [75]

    Citizenship and

    Torney-Purta, Judith and Lehmann, Rainer and Oswald, Hans and Schulz, Wolfram , year =. Citizenship and

  71. [76]

    Read, Tony , month = jul, year =. Where

  72. [77]

    AERA Open , author =

    Content. AERA Open , author =. 2020 , pages =. doi:10.1177/2332858420940312 , abstract =

    work page doi:10.1177/2332858420940312 2020

  73. [78]

    Open educational resources: education for the world? , volume =

    Richter, Thomas and McPherson, Maggie , month = aug, year =. Open educational resources: education for the world? , volume =. Distance Education , publisher =. doi:10.1080/01587919.2012.692068 , abstract =

    work page doi:10.1080/01587919.2012.692068 2012

  74. [79]

    RoBERTa: A Robustly Optimized BERT Pretraining Approach

    Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin , month = jul, year =. doi:10.48550/arXiv.1907.11692 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1907.11692 1907

  75. [80]

    , month = jan, year =

    Gozalo-Brizuela, Roberto and Garrido-Merchan, Eduardo C. , month = jan, year =. doi:10.48550/arXiv.2301.04655 , abstract =

    work page doi:10.48550/arxiv.2301.04655

  76. [81]

    Galactica: A Large Language Model for Science

    Taylor, Ross and Kardas, Marcin and Cucurull, Guillem and Scialom, Thomas and Hartshorn, Anthony and Saravia, Elvis and Poulton, Andrew and Kerkez, Viktor and Stojnic, Robert , month = nov, year =. Galactica:. doi:10.48550/arXiv.2211.09085 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2211.09085

  77. [82]

    Zhang, Chaoning and Zhang, Chenshuang and Li, Chenghao and Qiao, Yu and Zheng, Sheng and Dam, Sumit Kumar and Zhang, Mengchun and Kim, Jung Uk and Kim, Seong Tae and Choi, Jinwoo and Park, Gyeong-Moon and Bae, Sung-Ho and Lee, Lik-Hang and Hui, Pan and Kweon, In So and Hong, Choong Seon , month = apr, year =. One. doi:10.48550/arXiv.2304.06488 , abstract =

    work page doi:10.48550/arxiv.2304.06488

  78. [83]

    Borji, Ali , month = apr, year =. A. doi:10.48550/arXiv.2302.03494 , abstract =

    work page doi:10.48550/arxiv.2302.03494

  79. [84]

    Extracting

    Carlini, Nicholas and Tramèr, Florian and Wallace, Eric and Jagielski, Matthew and Herbert-Voss, Ariel and Lee, Katherine and Roberts, Adam and Brown, Tom and Song, Dawn and Erlingsson, Úlfar and Oprea, Alina and Raffel, Colin , year =. Extracting

  80. [85]

    Ethical and social risks of harm from Language Models

    Weidinger, Laura and Mellor, John and Rauh, Maribeth and Griffin, Conor and Uesato, Jonathan and Huang, Po-Sen and Cheng, Myra and Glaese, Mia and Balle, Borja and Kasirzadeh, Atoosa and Kenton, Zac and Brown, Sasha and Hawkins, Will and Stepleton, Tom and Biles, Courtney and Birhane, Abeba and Haas, Julia and Rimell, Laura and Hendricks, Lisa Anne and Is...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2112.04359

Showing first 80 references.