An official website of the United States government
Few studies have compared Large Language Models (LLMs) to traditional Machine Learning (ML)-based automated scoring methods in terms of accuracy, ethics, and economics. Using a corpus of 1000 expert-scored and interview-validated scientific explanations derived from the ACORNS instrument, this study employed three LLMs and the ML-based scoring engine, EvoGrader. We measured scoring reliability (percentage agreement, kappa, precision, recall, F1), processing time, and explored contextual factors like ethics and cost. Results showed that with very basic prompt engineering, ChatGPT-4o achieved the highest performance across LLMs. Proprietary LLMs outperformed open-weight LLMs for most concepts. GPT-4o achieved robust but less accurate scoring than EvoGrader (~500 additional scoring errors). Ethical concerns over data ownership, reliability, and replicability over time were LLM limitations. EvoGrader offered superior accuracy, reliability, and replicability, but required, in its development a large, high-quality, human-scored corpus, domain expertise, and restricted assessment items. These findings highlight the diversity of considerations that should be used when considering LLM and ML scoring in science education. Despite impressive LLM advances, ML approaches may remain valuable in some contexts, particularly those prioritizing precision, reliability, replicability, privacy, and controlled implementation.
more »
« less
spade: Synthesizing Data Quality Assertions for Large Language Model Pipelines
Large language models (LLMs) are being increasingly deployed as part of pipelines that repeatedly process or generate data of some sort. However, a common barrier to deployment are the frequent and often unpredictable errors that plague LLMs. Acknowledging the inevitability of these errors, we proposedata quality assertionsto identify when LLMs may be making mistakes. We present spade, a method for automatically synthesizing data quality assertions that identify bad LLM outputs. We make the observation that developers often identify data quality issues during prototyping prior to deployment, and attempt to address them by adding instructions to the LLM prompt over time. spade therefore analyzes histories of prompt versions over time to create candidate assertion functions and then selects a minimal set that fulfills both coverage and accuracy requirements. In testing across nine different real-world LLM pipelines, spade efficiently reduces the number of assertions by 14% and decreases false failures by 21% when compared to simpler baselines. spade has been deployed as an offering within LangSmith, LangChain's LLM pipeline hub, and has been used to generate data quality assertions for over 2000 pipelines across a spectrum of industries.
more »
« less
- Award ID(s):
- 2243822
- PAR ID:
- 10646608
- Publisher / Repository:
- ACM
- Date Published:
- Journal Name:
- Proceedings of the VLDB Endowment
- Volume:
- 17
- Issue:
- 12
- ISSN:
- 2150-8097
- Page Range / eLocation ID:
- 4173 to 4186
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Practitioners frequently take multiple samples from large language models (LLMs) to explore the distribution of completions induced by a given prompt. While individual samples can give high-quality results for given tasks, collectively there are no guarantees of the distribution over these samples induced by the generating LLM. In this paper, we empirically evaluate LLMs’ capabilities as distribution samplers. We identify core concepts and metrics underlying LLM-based sampling, including different sampling methodologies and prompting strategies. Using a set of controlled domains we evaluate the error and variance of the distributions induced by the LLM. We find that LLMs struggle to induce reasonable distributions over generated elements, suggesting that practitioners should more carefully consider the semantics and methodologies of sampling from LLMs.more » « less
-
Abstract BackgroundGenerative artificial intelligence (AI) large‐language models (LLMs) have significant potential as research tools. However, the broader implications of using these tools are still emerging. Few studies have explored using LLMs to generate data for qualitative engineering education research. Purpose/HypothesisWe explore the following questions: (i) What are the affordances and limitations of using LLMs to generate qualitative data in engineering education, and (ii) in what ways might these data reproduce and reinforce dominant cultural narratives in engineering education, including narratives of high stress? Design/MethodsWe analyzed similarities and differences between LLM‐generated conversational data (ChatGPT) and qualitative interviews with engineering faculty and undergraduate engineering students from multiple institutions. We identified patterns, affordances, limitations, and underlying biases in generated data. ResultsLLM‐generated content contained similar responses to interview content. Varying the prompt persona (e.g., demographic information) increased the response variety. When prompted for ways to decrease stress in engineering education, LLM responses more readily described opportunities for structural change, while participants' responses more often described personal changes. LLM data more frequently stereotyped a response than participants did, meaning that LLM responses lacked the nuance and variation that naturally occurs in interviews. ConclusionsLLMs may be a useful tool in brainstorming, for example, during protocol development and refinement. However, the bias present in the data indicates that care must be taken when engaging with LLMs to generate data. Specially trained LLMs that are based only on data from engineering education hold promise for future research.more » « less
-
Automating hardware design could obviate a signif-icant amount of human error from the engineering process and lead to fewer errors. Verilog is a popular hardware description language to model and design digital systems, thus generating Verilog code is a critical first step. Emerging large language models (LLMs) are able to write high-quality code in other programming languages. In this paper, we characterize the ability of LLMs to generate useful Verilog. For this, we fine-tune pre-trained LLMs on Verilog datasets collected from GitHub and Verilog textbooks. We construct an evaluation framework comprising test-benches for functional analysis and a flow to test the syntax of Verilog code generated in response to problems of varying difficulty. Our findings show that across our problem scenarios, the fine-tuning results in LLMs more capable of producing syntactically correct code (25.9% overall). Further, when analyzing functional correctness, a fine-tuned open-source CodeGen LLM can outperform the state-of-the-art commercial Codex LLM (6.5% overall). We release our training/evaluation scripts and LLM checkpoints as open source contributions.more » « less
-
Diversity in training data, architecture, and providers is assumed to mitigate homogeneity in LLMs. However, we lack empirical evidence on whether different LLMs differ \textit{meaningfully}. We conduct a large-scale empirical evaluation on over 350 LLMs overall, using two popular leaderboards and a resume-screening task. We find substantial correlation in model errors---on one leaderboard dataset, models agree 60% of the time when both models err. We identify factors driving model correlation, including shared architectures and providers. Crucially, however, larger and more accurate models have highly correlated errors, even with distinct architectures and providers. Finally, we show the effects of correlation in two downstream tasks: LLM-as-judge evaluation and hiring---the latter reflecting theoretical predictions regarding algorithmic monoculture.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.14778/3685800.3685835
