Align card with actual data: S1-S4, single-gap, dual schema, reachability

README.md

@@ -34,33 +34,62 @@ identification. Unlike monolithic recognition sets, these isolate
 *orchestration*: every item is built so that a strong single-image model must
 lose, and the deciding evidence is reachable only by *calling the right tool*.
 
-- **`agentic`** — 500 items, each with an engineered **information gap** (see
-  scenarios) that a tool-free model cannot close.
-- **`calibration`** — 400 items, stress-testing *when to commit* to a species
-  vs. degrade to genus/family or abstain.
+- **`agentic`** — **500** items, each with an engineered **information gap**
+  (scenarios S1–S4 below) that a tool-free model cannot close. Every item's
+  minimal necessary tool set (`mnts`, 1–2 tools) is hand-verified
+  (`mnts_confirmed = true`). v0 items are **single-gap** (one deciding tool);
+  multi-gap and fault-injection layers are left to a future release.
+- **`calibration`** — **400** items, testing *when to commit* to a species vs.
+  degrade to genus/family or abstain. Each item carries a **`reachability`**
+  label — whether the true species is actually within a tool's top-*k* — which
+  makes the **tool ceiling** explicit: **303/400 (76%)** are `truth_missing`
+  (no tool can surface the truth), the rest `top1_high` / `mid` / `top1_low`.
 
-Both are stratified by frequency tier (common / uncommon / rare) so the
-long-tail **tool ceiling** is visible rather than averaged away.
+Both sets are stratified by frequency tier (common / uncommon / rare):
+agentic ≈ 181 / 178 / 141, calibration ≈ 134 / 133 / 133.
 
-## Fields (per item)
+> **The two splits use different schemas** (the agentic set is scenario-based;
+> the calibration set is reachability-based) — see the field tables below.
+
+## Fields
+
+**Common to both splits:** `id` · `modality` (`image` / `sound` / `image_sound`)
+· `query` · `mnts` (minimal necessary tool set) · `freq_tier`
+(`common`/`uncommon`/`rare`) · `geo_band` · `true_species_ebird` (eBird code —
+**scoring only, not shown to the model**) · `image` (iNaturalist URL, `null`
+for sound-only) · `audio` (Xeno-canto URL, `null` for image-only).
+
+**`agentic` only:**
 
 | field | meaning |
 |---|---|
-| `id` | stable item id |
-| `image` | source image URL (iNaturalist open-data) — `null` for sound-only items |
-| `audio` | source audio URL (Xeno-canto) — `null` for image-only items |
-| `scenario` | S1 look-alike (voice-diagnostic) · S2 range/season prior · S3 poor quality · S4 multi-bird/clutter · S5 OOD/abstain |
-| `difficulty` | L1 single gap · L2 double gap · L3 gap + fault/OOD |
-| `modality` | `image` / `sound` / `image+sound` |
-| `query` | the natural-language question posed to the model |
-| `mnts` | **minimal necessary tool set** — the calls without which the item is information-theoretically unsolvable |
-| `mnts_confirmed` | whether MNTS was hand-verified |
-| `freq_tier` | `common` / `uncommon` / `rare` |
-| `geo_band` | coarse climate band |
-| `true_species_ebird` | ground-truth taxon (eBird code) — **scoring only, not shown to the model** |
+| `scenario` | S1 look-alike (voice-diagnostic) · S2 range/season prior · S3 poor image quality · S4 multi-bird / clutter |
+| `difficulty` | `L1` (single-gap; the only layer in v0) |
+| `mnts_confirmed` | MNTS hand-verified (always `true` here) |
 | `privileged_geo` | true `{lat, lon, date}` — **held out of the prompt**; the model must call `geo_prior` to use location |
 | `cf_geo` | a **counterfactual** `{lat, lon}` for the anti-memorization probe (a model that truly uses the tool changes its verdict) |
 
+**`calibration` only:**
+
+| field | meaning |
+|---|---|
+| `purpose` | `"calibration"` |
+| `reachability` | is the truth reachable through a tool? `top1_high` · `mid` · `top1_low` · `truth_missing` |
+| `lat`, `lon`, `date` | observation geo/date (scoring / tool input; keep out of the prompt) |
+
+## Scenarios (agentic set)
+
+| scenario | information gap | required orchestration |
+|---|---|---|
+| **S1** look-alike, voice-diagnostic | near-identical plumage (e.g. *Empidonax*) | `classify_sound` on the diagnostic call |
+| **S2** range/season prior | candidate looks right but is implausible here/now | `geo_prior(lat,lon,date)` to down-weight |
+| **S3** poor image quality | distant / blurry / backlit; field marks hidden | `quality_gate` → `zoom`/`enhance` → reclassify |
+| **S4** multi-bird / clutter | several birds or a tiny subject | `detect_bird` → `crop` → classify the isolate |
+
+Each scenario has 125 agentic items. **Abstention / out-of-distribution**
+behavior is tested by the separate **calibration** set (via `reachability` and
+the option to degrade or abstain), not as an agentic scenario.
+
 ## Protocol (important)
 
 - **Privileged information is never shown.** GPS, date, and the true taxon are
@@ -96,8 +125,7 @@ for line in open("agentic.jsonl"):
 
 Media licenses follow their sources (citizen-science; many are non-commercial),
 hence the `cc-by-nc-4.0` label for this annotation package. Taxonomy is on the
-eBird/Clements crosswalk. A small sample of pre-fetched images is not included to
-keep the package license-clean; the URLs above are the canonical source.
+eBird/Clements crosswalk.
 
 ## Citation