Research questions

feature · paper proposed

An earlier declaration was accepted by Zach Mainen; it has moved since — this is da2445194a34, unaccepted. · no paper read under it

What did this paper set out to answer, and which claim answers each question?

A capability that revises existing artifacts. It runs once, and then it is part of the schema.

Part of Tree — What does the paper assert, and how do its claims depend on each other?

How it works

Given the abstract (and the Introduction where the prepared paper carries it) and the paper's existing hypothesis and alt- claims — slug and sentence — it returns the questions the paper states and which claim answers which. A hypothesis is the answer the paper commits to; the alternatives it rules out are the other answers to the same question. It then writes questions: onto the paper's index.md and addresses: onto the named claim files, editing their frontmatter in place.

--dump-prompt writes the exact request and exits, and --answer feeds a reply back through the same validation — an invented slug or a question no claim addresses is dropped rather than written — so a model that has no backend here can still answer it.

Rests on

Feeds — a change here disturbs these

How it is defined

A model answers this layer, so the prompt is the layer. It is reproduced below from the committed file, and it is a declared input — editing it makes every run that used it stale.

extract/prompts/questions.mdthe prompt it runs undernot in the repository

The declaration names this path and the repository does not have it. An input that does not exist hashes to nothing, so it cannot make a run stale — the layer is declared to depend on something it is not in fact tracking.

extract/prompts/contract/vocabulary.mdthe prompt it runs undernot in the repository

The declaration names this path and the repository does not have it. An input that does not exist hashes to nothing, so it cannot make a run stale — the layer is declared to depend on something it is not in fact tracking.

What it produces3 questions

One paper, as the worked example — Gadeke, v6. Read from runs/gadeke-2026-guilt-insula/questions.output.json · 2 KB. model supplied:runs/gadeke-2026-guilt-insula/questions.answer.v6.json

# idtext
1 q1 Does the larger decrease in momentary happiness after a low outcome for the partner when the participant made the choice reflect responsibility-contingent interpersonal guilt, rather than general agen…
2 q2 Is the anterior insula, together with its condition- and choice-dependent connectivity to prefrontal cortex, the neural substrate of this interpersonal guilt, responding more when the participant is r…
3 q3 Does a neural substrate track the participant's responsibility for the partner's outcomes, representing partner reward prediction errors more strongly when they arise from the participant's own choice…

Across the corpus

9 not run · 1 stale·a paper links to its own cell, where this layer's output for it is rendered

Inputs and outputs

Reads, besides its dependencies
Produces
  • runs/{paper}/questions.output.json

One per paper — the table above links each one that exists.

Views
  • table — rendered above, over the 3 questions in the artifact

Running it

The command comes from the declaration, so this text and what actually runs cannot diverge. pipeline.py run also runs the unmet dependencies first.

python3 scripts/pipeline.py run <paper> questions

Underneath, that runs cd extract && python3 -m claim_graphs.cli questions --paper {paper}.