Consolidate StateBench to corrected dense-100m

Keep default and corrected dense-100m; remove legacy dense-max and grounded aliases.

README.md

@@ -32,8 +32,6 @@ configs:
 
 # StateBench v1
 
-> ⚠️ **Temporary quality notice:** `dense-100m` v1.3 has known prompt/oracle mismatches in six surface dialects. Do not start training on it. The audited v1.4 rebuild is in progress. `default` is unaffected.
-
 ### 110 million verified state episodes for pretraining, SFT, and on-policy SDFT
 
 Train models to **retain**, **edit**, **address**, and **transform** state across long execution traces.
@@ -45,7 +43,7 @@ Train models to **retain**, **edit**, **address**, and **transform** state acros
 | Configuration | Scale | Use when |
 |:--|:--|:--|
 | **default** | 10,000,000 programs · 7.80 GiB | You need 36 explicit families and varied surfaces |
-| **dense-100m** | 100M episodes · 96.60 GiB | You need high-throughput state training with grounded, structurally varied prompts |
+| **dense-100m** | 100M episodes · 98.61 GiB | You need high-throughput state training with grounded, structurally varied prompts |
 
 All configurations contain structured prompt, completion, and teacher-only privileged context. Every target is computed by an executable reference machine, never authored by a teacher model.
 
@@ -106,9 +104,9 @@ The default configuration includes a ready-to-use text column. Dense configurati
 | State systems | Transactional + temporal · Adaptive + control · Concurrent + distributed · Graph + structures |
 | Stress systems | Robust streaming · Multi-agent state · Algebraic automata · Memory pressure |
 
-default implements 36 explicit behavioral families. The corrected dense-100m configuration rematerializes 100M verified trajectories through 24 domain dialects: the 16 primitive domains and 8 compositions above. Its 32 document structures and 64 metamorphic surface passes are physically present in the prompts, including data formats, code-like records, logs, traces, conversations, and nested documents.
+default implements 36 explicit behavioral families. The corrected dense-100m v1.4 configuration rematerializes 100M verified trajectories through 24 domain dialects: the 16 primitive domains and 8 compositions above. Its 32 parseable document structures and 64 distinct metamorphic surface passes are physically present in the prompts, including data formats, code-like records, logs, traces, conversations, and nested documents.
 
-The dense semantic kernel trains long-horizon set, remove, additive, multiplicative, observation, and no-op transitions. Domain dialects diversify how those transitions are presented; they do not turn each label into a full simulator of that real-world system. StateBench remains synthetic state training, not a factual natural-language corpus.
+The dense semantic kernel trains long-horizon set, remove, additive, multiplicative, observation, and no-op transitions. Its 384 balanced `families` values are curriculum labels spanning domain and profile combinations, not 384 different VM mechanisms. Domain dialects diversify how transitions are presented; they do not turn each label into a full simulator of that real-world system. StateBench remains synthetic state training, not a factual natural-language corpus.
 
 ## Expected model capabilities
 
@@ -141,9 +139,9 @@ StateBench trains state handling, not factual knowledge. It does not by itself e
 
 | Configuration | Density | Token estimate |
 |:--|--:|--:|
-| dense-100m | 160.09 units / 1k chars | ~107–120B tokens |
+| dense-100m | 160.04 units / 1k chars | 106.6B GPT-2 · 118.9B Qwen3 |
 
-The token range was measured on 1,000 real packed rows with GPT-2 and Qwen3 tokenizers. Actual training cost depends on chat templates, truncation, packing, and the SDFT generation policy.
+The token estimates were measured on 1,000 v1.4 packed rows (seed 42). Actual training cost depends on chat templates, truncation, packing, and the SDFT generation policy.
 
 ## Integrity
 

data/dense-100m/train/part-00000.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b53f80ddf41167016724b00b81021a4285d94358015d9a498509b5da54ae39a7
-size 423333819

data/dense-100m/train/part-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e32ad6394a94323c66f06a174c6eaf62dbf65a9e32502b685e152998f40246c8
-size 423454676

data/dense-100m/train/part-00002.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:99c6525922a2c3bfafec376f87045b99fcfbbc7a9350225f548c973104450e40
-size 423566334

data/dense-100m/train/part-00003.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bd626889ed91d72226cca81317bf74a322ac1b07e572575ef1c4eac5f7c0e8ea
-size 423671197

data/dense-100m/train/part-00004.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:abc9b2787f8a936eb5e86811fd90c2c5fc15a267913fa68c468057bdcf3c4deb
-size 423611778

data/dense-100m/train/part-00005.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f13884f2b799b7170155a1c2d4ece937d469c8ca983245b264c3cec9ab22536a
-size 423674140

data/dense-100m/train/part-00006.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:23a3379fe993b6ffca5d63f5b33664db194eb0fe287ae3eaa7da3b6fd0c95e43
-size 423583036

data/dense-100m/train/part-00007.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:aa0623bc2e06088cdb70c008f0688ad8090389f3b872d5eda5a070044846379f
-size 423629493

data/dense-100m/train/part-00008.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:629ce81140980243d1a084f5ffe6de1abd7f2e7ad144a2d8ef0effe384471046
-size 423484369

data/dense-100m/train/part-00009.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1c780025dddc584ccb93f61244f1ced95643036892fd1c2f3d76d3a4a52b8289
-size 423587841

data/dense-100m/train/part-00010.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:de238393bd9c3f7f3cdda5bddf642c017f462b04146450bc8defcc7fd39763e2
-size 423422742

data/dense-100m/train/part-00011.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f71a8842a700ae96df4111ae36e0d64b68c48a6fbb02269f35aa276e0bf496d1
-size 423514659

data/dense-100m/train/part-00012.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b7fa43cb25324ea3272516f3765fd062cdff37fcf1d3dd7ac79affc94b230412
-size 423344700

data/dense-100m/train/part-00013.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:374784f28e55328daf2d77ff67d9f32567c68c6e9fbcee18647ec63215162f85
-size 423455038

data/dense-100m/train/part-00014.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:37bcd2eb19e411878546d5ca064a3278146930ac10c72e3fe0689d37c07c39db
-size 423479380

data/dense-100m/train/part-00015.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:482c6e10d3f5acabc62928aab133a95352a4c75b1e2bae69b534a5b015244f3e
-size 423716080

data/dense-100m/train/part-00016.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e1f52df90ab31df00216d02e4d5973801694356eb08b4a7571adb24fca79b54a
-size 423572402

data/dense-100m/train/part-00017.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:6a71cede9017003d0f82ae935c76d7d4852b8e9ce0ab369a0dab205eca4359a7
-size 423648605

data/dense-100m/train/part-00018.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1d35cb882b7edfbfdec9d8c0357a5e0a4d8a3a4414728d9e7568b0e46448fac2
-size 423598663

data/dense-100m/train/part-00019.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:33c0779244894f4a580e0103ccd4da75f7aa7e1dbf2a84a3b4c9cfb3f232c1b2
-size 423558887

data/dense-100m/train/part-00020.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1a8bc5fc9355c06d24f0c34b3d6c67cee358380527acc287915fab6b25e5d3e8
-size 423526846

data/dense-100m/train/part-00021.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:770c029625cdd0a3021067c0b18979e159e58bc75f9c84e8dd03bde4405841a3
-size 423571308

data/dense-100m/train/part-00022.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:fe2f27671f68a624465997cb232b7186bc593018a4401f82d34db5107d905d5e
-size 423527073

data/dense-100m/train/part-00023.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:346faae41159c7ba6f831598c53879e227e2cbf5da252c631005259be2176ae1
-size 423505661

data/dense-100m/train/part-00024.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:29f76513611347bda2d90c0ed8d91a9b3c64bcb5f7af3388cdde57dfee730c00
-size 423363320

data/dense-100m/train/part-00025.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c5e75ed4bd7c0c8559bcaf1cc61b2eec19983d1cb12ebd77368c142ad7872fff
-size 423439131

data/dense-100m/train/part-00026.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:40a67a5364a705d4bb782fece334815e44cc4d2d109c98e0c5bf906fe1c0cc7b
-size 423556948

data/dense-100m/train/part-00027.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f14abfb32467d5d8a55425f166bdc25ae7de0af0e7e05681a770820267b180ad
-size 423722828

data/dense-100m/train/part-00028.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:be46c5018e30a92c4f37ce2d3ada01472d0a0e3fa5a8c52e599d2839c6e08985
-size 423657904

data/dense-100m/train/part-00029.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:09f07c7f8d3884102b4ac6872f52918965c79e533f8293853ea931488af5c7d3
-size 423644486

data/dense-100m/train/part-00030.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3409a56151d4892973233c5a50fa2ebe9608bce0d3c0de57ee919d81a0cf3c1d
-size 423509856

data/dense-100m/train/part-00031.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2378de28ef1ee23163edf2538d4ed569b7458783afb5ffc62d11484bc298243e
-size 423555933

data/dense-100m/train/part-00032.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a404ecefddd11bf2be104684114396a1c29dfa7a4de74909c91301c963fda04d
-size 423539507

data/dense-100m/train/part-00033.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0bc699b770f280ae99ffb496781081e55de92963bfd55c50ae5e72920e4361cd
-size 423524121

data/dense-100m/train/part-00034.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:8b3d44bda7101e50e81cbc6ec309d07b8a56dc36501634ed24904c18a7aeef65
-size 423417120

data/dense-100m/train/part-00035.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5db991e04085fe4447141aa7e7b58651341400861de6bec4a9f555d54e6b27d4
-size 423501946

data/dense-100m/train/part-00036.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2a730f5687fb112a81918b35c915fa2a05008c175081afdfa0e6324548ad92d1
-size 423407051

data/dense-100m/train/part-00037.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e7f12e98ab4a4b6bc53cd8a324f0a81c4c3ce4ca21bb6173ff8195c3eb6a0ab3
-size 423440316

data/dense-100m/train/part-00038.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:925c9c233ac64e80cd774dbdd89c20583fb9f079953bebcfe0ca1ef0364503d7
-size 423562563

data/dense-100m/train/part-00039.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f635d4f1a70c9d3c31c86649cfdda32604d18f4743fa15c3fb6233be2db5893e
-size 423631046

data/dense-100m/train/part-00040.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:01840b206ad1d6946a89a3d75fdeabdc08348032a6e4c0d9eea60515c6bb5aab
-size 423697025

data/dense-100m/train/part-00041.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1a748006b21eaf41180a782d8d9b73a34df937cb3b938e5bb3385260fc9b9002
-size 423645986

data/dense-100m/train/part-00042.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:34e2ef9bd3761bdf0ab9c98b42aa888f2ec67c2106900b4de982b75af0b93e00
-size 423607757

data/dense-100m/train/part-00043.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7500ac59fde9c2cb7c3a5889ebfc0ef4cf8c27de963ac8428052b7d70b1b9e63
-size 423503420

data/dense-100m/train/part-00044.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:419855fc9cb81feeab62ac8145a00edf3f238c255ddae5ca489458fb56c09cb7
-size 423526066

data/dense-100m/train/part-00045.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9c7f300ff5b69ba2826234feffe9abf90a84b540254eeb9a9c871a4a7e4eb2d1
-size 423523715

data/dense-100m/train/part-00046.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9cb546b44d73b4d2a46c359aae36efbc5344d86c5b6cfc937e5e757d603c960e
-size 423555450

data/dense-100m/train/part-00047.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:82fef9318fb9bcac38bcd030003d6cc730fb59445d1b84d19698a1d4b1eb0815
-size 423490409

data/dense-100m/train/part-00048.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9188d79037bb11a71f117903785ac42ab8296187e294e03f3a506b5b73740f8a
-size 423438767