AI Paper Atlas

Research map / Large language models / Agents, code and math

Can Humans Teach Machines to Code?

Celine Hocquette, Johannes Langer, Andrew Cropper, Ute Schmid

AAAI 2026 · Large language models › Agents, code and math · approach: Code generation benchmarks · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use CS-educated instead of non-CS participants as teachershuman-provided examples · six list conceptspredictive accuracyNo clear effect
remove erroneous examples from participants' sets (ablation)NCS/CS human examples · dropkpredictive accuracy (%): NCS 92 -> 99; CS 66 -> 69Improved
remove erroneous examples from participants' sets (ablation)CS human examples · appendpredictive accuracy (%): 95 -> 100Improved
remove erroneous examples from participants' sets (ablation)NCS human examples · sortedpredictive accuracy (%): 85 -> 84No clear effect
give examples from an expert in program synthesis instead of non-expertssynthesis systems · all six conceptspredictive accuracy: expert 100% for every concept vs lower non-expert averagesImproved
sample examples randomly instead of using NCS human examplessynthesis systems · 2 examples, NCS grouppredictive accuracy: human significantly lower than uniform distribution (p=.01)Improved
sample examples randomly instead of using CS human examplessynthesis systems · 2 examples, CS grouppredictive accuracy: human significantly lower than both random distributions (p<.001)Improved
use RandomNormal instead of CS human examplessynthesis systems · 10 examples, CS grouppredictive accuracy: human lower than normal distributionImproved
use CS-trained instead of NCS participantshuman examples after error removal · dropkpredictive accuracy (%): NCS 99 vs CS 69 (p<.05)Hurt
use more examples (non-experts gave more than expert)human teaching · maxlistnumber of examples vs accuracy: NCS 6.7 and CS 6.6 examples vs expert 3No clear effect
use longer lists in examplesexpert examples · lastlearned concept: single length-6 example yields last instead of fourth elementImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy