Abstract ↔ claims

step · paper open

Declaration 98dd5c4dff2a has not been accepted by anyone. · no paper read under it

Which claims does the abstract carry, and which does it drop?

Mechanical: re-runnable from its declared inputs.

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

How it works

Cuts the abstract into sentences and asks of each which claims carry it, and of each claim whether any sentence of the abstract does. The output holds both directions, and the interesting column is the empty one: claims the tree holds that the abstract drops, and abstract sentences no claim accounts for.

An abstract is the paper's own summary of itself, so the difference between the two lists is the difference between what a paper found and what it advertises.

--dump-prompt writes the exact request and exits, and --answer feeds a reply back through the same validation — an unknown slug or a sentence number the abstract does not have is dropped, and the two directions are derived from one per-sentence mapping so they cannot disagree — so a model with 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/abstract-map.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 produces

One paper, as the worked example — Artiushin, v1. Read from site/src/data/abstract-mapping/artiushin-2026-spider-atlas.json · 5 KB. paperSlug artiushin-2026-spider-atlas

abstract · 1.4k characters
Spider orb-web building is a captivating, rare example of animal construction, whose neural underpinnings remain undiscovered. We created a three-dimensional atlas for the hackled-orb weaver, Uloborus diversus, based on immunostaining for the presynaptic component, synapsin, in whole-mounted spider synganglia. Whereas spider neuroanatomy has thus far been most comprehensively studied in cursorial species, this optically sectioned atlas contributes a continuous, finely resolved model of the central nervous system of an orb-web building spider. Aligned to this volume, we examined the expression patterns of neuronal populations representing many of the classical neurotransmitters and neuromodulators (GABA, acetylcholine, serotonin, and octopamine/tyramine), as well as a subset of neuropeptides (allatostatin A, crustacean cardioactive peptide (CCAP), FMRFamide, proctolin) – detailing immunoreactivity in an unbiased fashion throughout the synganglion to reveal co-expression in known structures (such as the arcuate body), as well as novel neuropils not readily apparent in prior spider research, including the tonsillar neuropil as well as a potential protocerebral bridge. These structures provide targets for future functional studies, and taken together, could represent a spider equivalent of the central complex, contributing to behaviors such as web-building.

Across the corpus

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

PaperStateVersionLast runOutputCell
A three-dimensional immunofluorescence atlas of the …backfilled from the artifact on diskrun not observedv12026-04-20artiushin-2026-spider-atlas.jsonjson
Distinct representational properties of cues and con…backfilled from the artifact on diskrun not observedv12026-04-20bouyeure-2026-fear-rsa.jsonjson
Computational modelling identifies key determinants …backfilled from the artifact on diskrun not observedv12026-04-20ejdrup-2026-dopamine.jsonjson
Contributions of insula and superior temporal sulcus…re-run for the current treestalev52026-09-12gadeke-2026-guilt-insula.jsonjson
Spatially targeted inhibitory rhythms differentially…backfilled from the artifact on diskrun not observedv12026-04-19headley-2026-inhibitory-rhythms.jsonjson
Feedback of peripheral saccade targets to early fove…backfilled from the artifact on diskrun not observedv12026-04-19kammer-2026-foveal-feedback.jsonjson
iGABASnFR2 is an improved genetically encoded protei…backfilled from the artifact on diskrun not observedv12026-04-20kolb-2026-igabasnfr2.jsonjson
A deep learning pipeline for mapping in situ network…backfilled from the artifact on diskrun not observedv12026-04-20rozak-2026-neurovascular-dl.jsonjson
Self-association enhances early attentional selectio…backfilled from the artifact on diskrun not observedv12026-04-20scheller-2026-self-prioritization.jsonjson
Impaired excitability of fast-spiking neurons in a n…backfilled from the artifact on diskrun not observedv12026-04-20wengert-2026-kcnc1.jsonjson

Inputs and outputs

Reads, besides its dependencies
Produces
  • site/src/data/abstract-mapping/{paper}.json

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

Views
  • document — rendered above, from the artifact itself
  • comparison — on the cell page, two versions aligned by the matcher, wherever the ledger holds more than one

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> abstract-map

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