diff --git a/.gitattributes b/.gitattributes
index 99d6df39051d941f2e301269c43c2c3a3465f555..de3b821f3a9766d8ba7203844b8e2271428f5214 100644
--- a/.gitattributes
+++ b/.gitattributes
@@ -1024,3 +1024,72 @@ atpo_hotpotqa_qwen4b/20260506_102808/rollout/231.jsonl filter=lfs diff=lfs merge
 atpo_hotpotqa_qwen4b/20260506_102808/rollout/240.jsonl filter=lfs diff=lfs merge=lfs -text
 atpo_hotpotqa_qwen4b/20260506_102808/rollout/222.jsonl filter=lfs diff=lfs merge=lfs -text
 atpo_hotpotqa_qwen4b/20260506_102808/rollout/220.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/210.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/204.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/245.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/247.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/214.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/251.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/207.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/225.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/236.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/226.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/224.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/217.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/229.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/202.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/203.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/243.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/241.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/212.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/234.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/228.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/201.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/227.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/242.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/244.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/230.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/248.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/239.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/211.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/218.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/219.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/216.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/205.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/213.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/206.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/233.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/250.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/249.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/209.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/237.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/238.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/215.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/223.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/222.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/208.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/210.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/226.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/228.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/207.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen4b/20260506_102808/rollout/221.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/225.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/214.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/204.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/224.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/203.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/202.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/227.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/220.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/205.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/215.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/208.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/221.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/201.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/211.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/206.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/209.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/217.jsonl filter=lfs diff=lfs merge=lfs -text
+atpo_hotpotqa_qwen8b/20260506_124901/rollout/212.jsonl filter=lfs diff=lfs merge=lfs -text
+cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/181.jsonl filter=lfs diff=lfs merge=lfs -text
+cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/146.jsonl filter=lfs diff=lfs merge=lfs -text
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/global_step_250/data.pt b/atpo_hotpotqa_qwen4b/20260506_102808/global_step_250/data.pt
new file mode 100644
index 0000000000000000000000000000000000000000..69724db5a40c42c0cdd26777eb51bd629bcd8e1e
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/global_step_250/data.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8810ecd4df9a2337ed2aceb01e57f5b03e4ca7dd3c7c5d614d0b7a3fe0959e5a
+size 1492
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/201.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/201.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..5f1af942f9d2975c6eedca329e5e371b09ba1582
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/201.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf7458bf89768088ef8d01c3a81833f53958a06dd9e994025e7c0cc10cfa954d
+size 13721379
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/202.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/202.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..c5d61be446e66c919f0ec13033679d03deb3bd20
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/202.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08926dcddcb0019b3a84c5aa3074f716425deedea1cbb7182e4960e182381372
+size 14083279
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/203.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/203.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..f89fa68e985dd9c38b3b40a0a20425036d949b2a
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/203.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44ba7def4aa5cd2b7339ece9c75ddb12aa64537b19f29a676ddb2946830e14ac
+size 14509445
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/204.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/204.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..0fa267173de6b64244430efc6339c2c722b165e2
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/204.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e7c20c185a3a40760cb8066e125a4eb7b5a434370f124094d84d334140ef2af
+size 15109988
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/205.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/205.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..b4406e599edc34f3703a4bb017d788c898f9a54e
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/205.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b86478539b9ecca6867ee8d67f5d7d7bf413f304c52bb6cc8c8e4cf2d44e31c9
+size 15380249
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/206.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/206.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..4139ac210bcd0d8fc87fff48c69aa6f1ca90f1f1
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/206.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cc69d5f933400e4cb69169a2c44a2227c1e5bd8dbc658feb494076ba832715d
+size 14224659
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/207.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/207.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..a83fd31932fde0034235b8ef4ab0aae11d7ad759
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/207.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:249e22965714ca750fe5bbeb06bfd480f592b4ff5d40a974d41f0eac4895d699
+size 15247917
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/208.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/208.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..e8c00612c2e7362f7c618eaab5dc0e730366a8ae
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/208.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b202c81ffb4c4ba9b192abeec5ae033b9425b2f888b85ee52ece831d189f790
+size 13990863
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/209.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/209.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..23767accde293fbab2aaffdc54a448ff97077cee
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/209.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:faba9924b986ac55860bce290b9c5f1c726ac18da92caddd1cbff1822ee69023
+size 14852027
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/210.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/210.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..c297b3788acb19cf3dfb5e353288d62e455e44da
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/210.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e61316dd24b52bdba2619571438cdd636434914430f1d2ea5a05eabfa1f66d8
+size 15185584
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/211.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/211.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..49a18a453aac5a382068c9243a285da1571c8ca5
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/211.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:943c7c0cdb185d76c5b378753108c9c3656e956e5482542a5cadcd92910adab1
+size 15625519
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/212.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/212.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..a94a1353276909444d58beb3645b256f258fcec9
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/212.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e0fca69f171f63dee49a37acc954ed75bf382aafc73eff99a9534fffcb649bd
+size 13688858
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/213.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/213.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..5f1e0af7b526edd6f5936de649071815591ed69a
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/213.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff9f7a18c4975a61bfa172b61abd14e47ae81df27605a3c751505afdc2321db4
+size 15799598
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/214.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/214.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..58342d1b298974956eeb2ef82bcd802e54186f6d
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/214.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d89fe0ceef6d2ed9a90f9bcbee9881407fd7d5a8d954f9a16a97c78e73034d5
+size 14406996
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/215.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/215.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..fa28f6ed904f4fa4208a2e9ea8d5e1ed208e0450
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/215.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7663a7e70ef3dd84b498980dd2f13bcc393155de077378934ea8a405d7164424
+size 14106854
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/216.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/216.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..9084dd3e239f68cdafe04bce03b4f4213f5e7351
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/216.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72e35111db6d0a2c0d436bb096c61da24fb8896a876537cad90e9ea6a41cedd9
+size 14040850
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/217.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/217.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..f03ad20dc3505d84639a78e280882107f3739e19
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/217.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:098703cf179628596b10269875e61b6c9f2a48d8ab3dc6d33f6f09ff70f6ee7f
+size 14067389
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/218.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/218.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..dd7ad4303e4913ba4a538b74a8c2ea3250580f36
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/218.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc6b24f4594f2b32b6dc28496f97a939b23213aa426b8b3bea9b86263dcda843
+size 15113874
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/219.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/219.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..7ef248d6fbe6e6c2ce0817a57a2011cf02a81620
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/219.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fd35cc93b8b03eb8451fd5e8987af08291228a548eca43d0d1ea79a9c35685d
+size 13311290
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/221.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/221.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..bbc187209ec81b2fd428d75dc2a77687a9809290
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/221.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3bfe8c8577a605edbdcdbbe3d37131423a44e70e6f07015baf6eb0b31dc2067e
+size 14264146
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/224.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/224.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..a39f4b8e20e24a33d4dcabc1659170287f164bd6
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/224.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e43e4e6a15af6d0f5dc2dba38c0645d2bb80ab282ba6ca11529f03d114a2aa0
+size 13602444
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/225.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/225.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..680e41be1228597ca5d1a6f190b3e039522a31f8
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/225.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4bc1c02812c3444c33ae9d9354ac01a57e8682d1a96caee10c941a8fdae36bf6
+size 13662906
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/226.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/226.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..5fcce3e163b18094ef3a98df2e42a79aaf66e0cc
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/226.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37a9413a4515ebc5f150e11401e9130d12a4c90c0e50c5ba79f5a392799943a4
+size 14291020
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/227.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/227.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..3a9c15027bf9b2e2ba2ef4911a0161fb111653df
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/227.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97f6061317a08c67a530fed3c09f105ad292d53398ad0bad3615c6921e8b21fd
+size 13772889
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/228.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/228.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..1bddbe01e40220e1f92dcef87b90d51dca9e34bb
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/228.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1ce2b5401e2074182e59df9bd926b8554b39faf889fd0cab184bd3fd167af3b
+size 14529028
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/229.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/229.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..ad1be871f9302394c3d8a6949845da4cda817137
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/229.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:170f7d3ab861e9e1695c1c41e8885da8425a50c02cf21438f263642f8f71f377
+size 13486765
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/230.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/230.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..4c24e201975cc5c445d6428d0f4361891946f3a5
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/230.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55a6288eaefc5ff9795538d22464197f3bdd1d9af527225328e46b425447cbae
+size 13750224
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/233.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/233.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..e457271100da1d151ad5d943df0b7b6d279ac3c0
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/233.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ffae9dad277cc8936f1abe85bdca852382841bc2ea08a8abb9dc19d4c27d15b7
+size 14414978
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/234.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/234.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..a195219a98d302be3097a3e443202129873218d0
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/234.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e962750f19bf323fc720d4bfd5ed577706a4dca3a90b6c7d30647f6bc53260d
+size 14485625
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/236.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/236.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..fd7bba0d13988e2ce21059a7c80821a6bfd2e541
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/236.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca55c063a2ee577845554f2add1226dea44d9cffc721b8f8484f6dc7a240d019
+size 14770953
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/237.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/237.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..96b8137f2f171f365f9430bbe7550589126a0af9
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/237.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c89f0f003c0852f9cf3cadd558a429a107627ba57c8fff3ad91f0978cdda1a1b
+size 14428402
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/238.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/238.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..83c96cd2e826ecba39114f69697b16a6e672d0e5
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/238.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:002b997ad07d42829c1968398278dcdcb4f2d255d439771b662f2c0a68dc7876
+size 14088600
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/239.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/239.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..6f1edad18b83d85f143decf51a73e6214c282ad1
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/239.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ffc5882856912adc872c12d2640c12af81eb8cedeb8e102a3d9fdd486b2d878
+size 14883315
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/241.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/241.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..8d7729fc0c1d42bb5f619d4bd3bf99602201db19
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/241.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c517080dcd487bcb9857e1c00660307782d64e4e109bf137352461197ec57844
+size 13926766
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/242.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/242.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..2275e7259dfcf07b19a5b702a2c727b72879699e
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/242.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c4ec4077f2b1cfce34ca29c4593725a0e2fccf7f64f54e67abf4e3ceebe2cb9
+size 14187218
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/243.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/243.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..7812d132487634970172e4509803f47d876b4a74
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/243.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e92daddfd92728c3bedf1c585ebd7c8647e889019dd7eacac2579657a1bd233f
+size 14595738
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/244.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/244.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..362d0e48f86925d5c7c70f7f2a1bd3cfc4e121c4
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/244.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcd4ec6e128126e105f3c657ae22d350648b38f76d457dcad005ec00a15fcee1
+size 15077015
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/245.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/245.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..d528d80ad2682927555686034c6231f346c7b789
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/245.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d764affa67ef360d8ffbdc4da1160e51b76410038fc92ebc198e14321fa6e2d
+size 14958159
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/247.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/247.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..529561f3302d511ed47263f7fd9b5cd74b1b06d7
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/247.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:819ef57dc1f7112c51dca9940b781687ceaab1ceb940ae4b7979924a9fa535e8
+size 15257394
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/248.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/248.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..66705a34aa18c3e7d203146dbcd4a5cd27caa68c
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/248.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f5aee53abd3af0ff3fd9786e365e7b87ecf5cc6f620e34bab28cd1dc2986f26
+size 15548435
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/249.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/249.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..13662b9d5217a660578173d78dd21a90698434ec
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/249.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be88a53d7b13cc613dcd64da3624f568453ad29f675e6f229ae917e49c715aa6
+size 14692575
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/250.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/250.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..4ac24c445b68846e10e8a284ebee86d0a6ad28d7
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/250.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d84c8a0d22fc9fc7669788928b299bfb887d1f11d0b8532f45064bfb8c8f627d
+size 15541975
diff --git a/atpo_hotpotqa_qwen4b/20260506_102808/rollout/251.jsonl b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/251.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..a87347209dc84b8944f322cf86dba18e9e9770da
--- /dev/null
+++ b/atpo_hotpotqa_qwen4b/20260506_102808/rollout/251.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9de8c9cbb3339e140db2d2da326797bdfa13e52ee2a9509d69af04ec0a964092
+size 15233597
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/201.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/201.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..d476fb5a8fcffb66bdebab5b5b1c98bd9d948a9f
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/201.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2317030deeb59e2f8761ae89e1f839fd94f381446923bb52b46ea74a80f92ea3
+size 13439044
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/202.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/202.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..3fdda49ee99dc7f2002e0a681a2d6d8684c0af85
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/202.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f626dbdedadeefa9faf62ca47d3d0abeab8351c0c7ad7013cca5e11131ff1b4
+size 13443707
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/203.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/203.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..8666eda04945da280a2412953316067389f6978b
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/203.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e0deb080e5ed9391d38b339e08da13a3b3ec35171cff05609111005bd9f5d4c
+size 13932764
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/204.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/204.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..ab4e114ce882eeb96de8a676ed58b2aea9316c8a
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/204.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:456d1ab0c70abc12615b640c4d276358c699a007a16c1ac0498c3a345df5b2c5
+size 14649132
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/205.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/205.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..3b2e1ade2482a8948deeec05dd29a98d041c74c4
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/205.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0ce7bf0fdd35f55e107a28e0f74eea7679cb1f5d4c3a7d171fec64ba6363972
+size 14122944
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/206.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/206.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..6407064529cfd5e85251c76d946b0c517dd3b70a
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/206.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0c7bb50402e66786f5d5aece5a27de46b89ded0933fa861104ccd5481ce572a
+size 13037224
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/207.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/207.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..18d284bb1a5b958e773cc2278435870d20ecdd1f
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/207.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c174dc94022c502d678bb6f81d8d0f578b16cb72877035e8abcfb366a0b450bb
+size 14266537
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/208.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/208.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..9be971af40719fccdb40d62ca5f990045cffcf48
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/208.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8c9eca2d1b98e1def9da2348ffe02047da064db9783368540c3277013f6e6d5
+size 13357216
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/209.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/209.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..396c16c79ffe6c008abc0a87374b75f55028d206
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/209.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df17b351d03aa220edd920957c7fef6389aa6890e435147efd60aa4baa2c8465
+size 13826324
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/210.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/210.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..b586403dbb6368c36402adca91a43a0e5c9cab4a
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/210.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f970914f8ed0ed50c82f8e24a92eccb5d8372a6b3f7d78c2ce40c76469ef2acc
+size 14234027
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/211.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/211.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..0dc303d2998b401ae5adb9b161d8f1d46f2fb304
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/211.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4f4544533f7a7f1aa82c0637021dfc63a461d79a604ee615cb27f02855f352b
+size 15008666
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/212.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/212.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..325a133d61904f2ae574f30850f30cec455ecfb9
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/212.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3997cf5a6434195cbd68651af16e469495742181725cf0939c4b84621d5738f
+size 14187944
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/214.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/214.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..5a10705857fd46a7ae1f446d4365eb79d6ac567c
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/214.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:424b8967761b697a4cd4291b397b8136084c7d73595351a0c11f056721e09b6a
+size 14401499
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/215.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/215.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..fe1c0b88fe3ad2acd16bd5b47fec453707b73e59
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/215.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f42db3321bf7b7cd02641bb746f71e83fe7e7ea6d4b55240d92d520b954ddc8d
+size 14703334
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/217.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/217.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..818c68b852836f2a88640f09622ad5b85870a856
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/217.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4003dde7f9fe33759c452f41af4bc14c49e91c54482e88cfe0eb94a7277410c9
+size 14766236
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/220.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/220.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..c3aca19952a4d808bfe349684b4e071da93a7e30
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/220.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3c07e850bf1abe95638a989e16c4d8ac515817e77c768e903473aa650f443f9
+size 15040062
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/221.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/221.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..2a2d7de5dfef36cf6358a303cdbec6f550936b5e
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/221.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99d8588ceb570d9eee8e1d868f4e124f57c8f928b0a94a6e7d6cee4cdabf3e96
+size 15697660
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/222.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/222.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..91f67e3802d4aa27ff9cac93f54e201b2b1b117a
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/222.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a12beaa07cd8fdabaf6dbdc2c261ac298512dd1322d24053ccff32ed26aae1b6
+size 14351969
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/223.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/223.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..d79290506efc2b012817072bce5eed81f6a8297b
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/223.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fb875231efc25f754c4775cc25d6d2404203a596cce36b68e96c3635408bd9a
+size 14909534
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/224.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/224.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..1f6ad4db3a82a581881751fffae54473c5b0bae7
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/224.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:994a63d6ea7a569ef0e8c69d798c14537e90bf82e7cfac8cc9a9ae7572d02eb4
+size 14756143
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/225.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/225.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..788e0dce8dd712cd54daf6f927ebe580c2e2e0ef
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/225.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da8f1e8b029c8d3b1892ca04341becf16115da21bf879a80d757aca37ca91599
+size 14646844
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/226.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/226.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..b64908d08c774be7c71d82a5e5cd3b219a8b0b99
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/226.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0a9eb4caf0fbebf4d4736962eeb3186f9e9261f5b2286262eccbe12e6f12e7b
+size 14712354
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/227.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/227.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..640789d795a0b9f4c9ae323a34d27309a9960153
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/227.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd2e98ad60d35bb958afe20fe63415ee1e779700036759a71fc95a6979b2ae59
+size 13786448
diff --git a/atpo_hotpotqa_qwen8b/20260506_124901/rollout/228.jsonl b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/228.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..c8a43095baef19dc2fb59ca0327a619a03818dc2
--- /dev/null
+++ b/atpo_hotpotqa_qwen8b/20260506_124901/rollout/228.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:baaececc4bbe88c02f47ee1fdd81b7d7832128698a14db6e36863c67e044e466
+size 15213827
diff --git a/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/global_step_200/actor/extra_state_world_size_8_rank_2.pt b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/global_step_200/actor/extra_state_world_size_8_rank_2.pt
new file mode 100644
index 0000000000000000000000000000000000000000..4fc618112e3c6a372c526c64c61ab3c7da5d6590
--- /dev/null
+++ b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/global_step_200/actor/extra_state_world_size_8_rank_2.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8740030293da6bc02a6c4128f94c5ce2e123aa098b1da4506b677c5d1b6093c
+size 14632
diff --git a/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/global_step_200/actor/extra_state_world_size_8_rank_6.pt b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/global_step_200/actor/extra_state_world_size_8_rank_6.pt
new file mode 100644
index 0000000000000000000000000000000000000000..4446e4ef03cec3fd1261b6dfce65046f6e71a590
--- /dev/null
+++ b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/global_step_200/actor/extra_state_world_size_8_rank_6.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c323cf5f0a110b84c4d42e1a025ceccee53ab38319cb09740570504d316eff3
+size 14632
diff --git a/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/global_step_200/data.pt b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/global_step_200/data.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a645dc6e0cd0f2bbfb7930a97f7dacecb69c780f
--- /dev/null
+++ b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/global_step_200/data.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fcb24549effe73885ded12714c61db3d2ea9f5df07953441eab16013c7b3d67a
+size 1492
diff --git a/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/146.jsonl b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/146.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..d7031faabb5497f28c79e21f72e72a117c6cd856
--- /dev/null
+++ b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/146.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f69ea2a273e1cd018c45f6189c1a9a1a79793d616a69ecec9f9006d0d22ca0f7
+size 13060318
diff --git a/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/181.jsonl b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/181.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..140332704436a81ef27e4faa93759719efbbf154
--- /dev/null
+++ b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/181.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8bbdb10714da95a9f650d34adc344e2ab0d192b600576021631af527bf177fbc
+size 16264186
diff --git a/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/63.jsonl b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/63.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..0e77817f650a70270007d62b57d9cc800c9be040
--- /dev/null
+++ b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/63.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44f27f13818484836d1f63927b1a30c3faa696b74dca1d8674ba603c46eb8591
+size 14109354
diff --git a/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/68.jsonl b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/68.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..80c53493857fc8839d3bf0ca134e410bcbbe1e67
--- /dev/null
+++ b/cache_hotpotqa_qwen4b_ablation_noadvantage/20260504_231531/rollout/68.jsonl
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b331b59d882e72f9b796f4920d8bbba928c49234c9103833a79929a006d7be4
+size 15054605