Zenaique

Match instruction dataset to its source / construction method

Match pairs·Medium·4.0 · 0·~2 min·Asked atCognizantKrutrimSwiggy·Relevant atCohereDatabricksMetaMistral AI
Attempt it

Drag each answer to line up with its matching prompt

LIMA

Community sourced multi-turn conversations with human review and ranking, released by LAION.

Alpaca

~15k pairs hand written by Databricks employees across categorised task buckets.

Dolly v2

52k self-instruct examples generated by an LLM (text-davinci-003) from a small seed set.

FLAN

Scraped logs of user conversations with ChatGPT, popular as a base for early open source instruction tuning.

OpenAssistant

Templated conversion of many classic supervised NLP datasets into instruction format at large scale.

ShareGPT

1000 hand curated examples from Meta, designed to show that less but better beats more but noisy.

TL;DR

Sort each dataset by who wrote the words and how: hand-curated (LIMA, Dolly), LLM-generated (Alpaca), templated benchmarks (FLAN), community multi-turn (OpenAssistant), scraped logs (ShareGPT).

Memory aid
Sign in to see the mnemonic that makes this stick.
Easy to grasp

Imagine six cookbooks. LIMA is 1000 recipes a chef perfected by hand to prove quality beats quantity. Dolly is recipes thousands of office workers wrote during a fun company drive. Alpaca had a robot churn out 52000 recipes cheaply from a few seed ideas. FLAN took thousands of old exam questions and reworded them all into one recipe format. OpenAssistant is a wiki where volunteers wrote conversations and voted on the best replies. ShareGPT is screenshots people posted of their chats with a famous robot, scraped into a pile. Knowing who cooked each one tells you why it tastes the way it does.

Concept explanation~2 min read

Everything you need to truly understand this topic: intuition, mechanics, step by step explanation, code, formulas, and worked example. Click to expand.

This question looks like trivia, but it is really a data-quality probe. Every instruction dataset is defined less by the topics inside it than by HOW it was constructed, and the construction method predicts the dataset's strengths, its biases, and its failure modes. An interviewer who asks you to match six famous datasets to their sourcing stories is checking whether you can reason about provenance, not whether you crammed a flashcard.

The clean way to hold all six in your head is a single grid with three columns: who authored the examples, at what scale, and by what mechanism. The author can be a human, an LLM, or a pile of pre-existing data. The mechanism follows from that, hand curation, self-instruct generation, templated conversion, or scraping. Once you sort the six along that axis, every pair becomes obvious and the distractors stop fooling you.

This deep dive walks the three mechanisms in turn, places each of the six datasets on the grid, and then draws the senior lesson the whole exercise is built around: curated demonstrations beat raw volume, which is exactly why LIMA exists as a deliberate counterpoint to Alpaca.

The three sourcing mechanisms

Every instruction dataset is built by one of three mechanisms, and naming the mechanism is half the battle. The mechanism is not a cosmetic label. It determines the cost per example, the upper bound on scale, the diversity of phrasing, and the legal status of the resulting weights.

The first is human authoring. People sit down and write (instruction, response) pairs by hand. This is the most expensive route and the highest quality per example, but it caps out at thousands rather than millions of examples. The signal is dense because a human writer brings judgement to every line, yet coverage is narrow because no team can hand-write a million prompts. LIMA and Dolly v2 live here.

The second is LLM generation. You give a teacher model a small seed set and let it synthesise thousands of new examples. Self-instruct is the canonical recipe: seed prompts go in, the teacher expands them into new instruction-response pairs, and a light filter removes near-duplicates. It collapses cost dramatically, but the output inherits the teacher's errors, its style, its refusal habits, and a licence that may forbid commercial use. Alpaca is the textbook case.

The third is reuse of existing data, in two flavours. Templating converts pre-existing supervised benchmarks into instruction phrasing at scale, which is the FLAN approach. Scraping harvests conversations that already happened, which is the ShareGPT approach. OpenAssistant is a fourth, crowd-sourced flavour where a volunteer community both writes the conversations and ranks competing responses, which makes it richer than a plain pile of pairs.

The human-written pair: LIMA and Dolly v2
The LLM-generated and templated giants: Alpaca and FLAN
The crowd-sourced pair: OpenAssistant and ShareGPT
Why provenance is the senior lesson
Sign in to unlock the full deep dive.

Situations where this technique stops working.

Sign in to see when this approach fails.

2–4 min · Everything important, quickly.

Sign in to see the quick scan of the deep dive.
DatasetWho wrote itScaleMechanism
LIMAMeta researchers1kHand-curated, quality over quantity
AlpacaAn LLM teacher52kSelf-instruct from seed set
Dolly v2Databricks employees~15kHand-written, categorised buckets
FLANExisting benchmarksVery largeTemplated conversion of supervised data
OpenAssistantVolunteer communityTens of thousandsMulti-turn dialogue with ranking
ShareGPTChatGPT usersTens of thousandsScraped conversation logs

Real products, models, and research that use this idea.

  • Stanford Alpaca kicked off the 2023 open-instruction wave by self-instructing 52k examples for roughly 500 USD, spawning Vicuna, Koala, and dozens of clones.
  • Databricks shipped Dolly v2 with a fully commercial-use licence precisely because the 15k pairs were employee-written, not distilled from a closed model.
Sign in to see more production examples.

What an interviewer would ask next. Try answering before peeking at the approach.

QWhy did Alpaca, despite 52k examples, often lose to LIMA's 1000 on alignment quality?
A

Frame around the self-instruct pipeline inheriting the teacher's errors and low diversity, versus hand curation maximising per-example signal. Quality density beats raw count once the model already has capability from pretraining.

2 more follow-ups an interviewer would ask next. Sign in to reveal them.

Red flags & common mistakes

The phrases that signal junior thinking. Click to expand.

Most common mistake

Mixing up Alpaca and LIMA. Both feel small and clever, but Alpaca is 52k cheap LLM-generated examples while LIMA is 1000 obsessively hand-curated ones with the opposite thesis.

Sign in to see all red flags and common mistakes.

60 second bullets to scan on the way to the call.

  • Three sourcing mechanisms: human-written, LLM-generated, templated or scraped

  • LIMA's less is more thesis and its 1000 example scale

Sign in to unlock the revision sheet.

Primary sources. Browse if you want the original framing.

Similar questions

Same topic, related formats. Practice these next.

4 curated
Next question
What is RLHF, and why is it used after pretraining?
MCQ·Easy