MétaCan
Menu
Back to cohort

Fine-Tuning Optimization of Small Language Models: A Novel Graph-Theoretical Approach for Efficient Prompt Engineering

2024· article· en· W4401164143 on OpenAlex
Venkata Gadiraju, Hao-Yu Tsai, Hsiao‐Chun Wu, Manali Singha, Chun-Yang Huang, Guannan Liu, Shih Yu Chang, Yiyan Wu

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

affAt least one author lists a Canadian institution in the pinned OpenAlex snapshot.

Bibliographic record

Venuenot available
Typearticle
Languageen
FieldBiochemistry, Genetics and Molecular Biology
TopicDNA and Biological Computing
Canadian institutionsCommunications Research Centre Canada
Fundersnot available
KeywordsComputer scienceGraphTheoretical computer scienceProgramming language

Abstract

fetched live from OpenAlex

In the realm of fine-tuning pre-trained language models in modern prompt engineering, we introduce a novel graph-theoretical approach to address the resource-intensive challenges to the conventional data fine-tuning methods for prompt engineering. Leveraging semantic and contextual prompt relationships, we propose to form a novel prompt graph, which facilitates a new comprehensive representation of prompt similarities. Building upon this new graph structure, our proposed approach can minimize the training time during the fine-tuning process for small language models by identifying and utilizing cliques corresponding to condensed subsets of highly similar prompts. This new strategic reduction in training data can greatly reduce the training time, particularly for resource-constrained applications in practice. Our proposed new approach leads to a significant reduction in the original prompt-graph order and a more focused and streamlined fine-tuning process. This data-reduction strategy demonstrates the potential to enable finetuning language models for prompt engineering with smaller datasets subject to less computational resource. The real run-time analysis for the training process of a small language model GPT2 have been undertaken to show the advantage of our proposed new approach.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.000
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: Simulation or modeling
GenreCandidate signal: Methods · Consensus signal: none
Teacher disagreement score0.647
Threshold uncertainty score0.276

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0000.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.000
Science and technology studies0.0000.000
Scholarly communication0.0000.000
Open science0.0000.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.016
GPT teacher head0.227
Teacher spread0.211 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Quick stats

Citations0
Published2024
Admission routes1
Has abstractyes

Explore more

Same topicDNA and Biological ComputingFrench-language works237,207