Add two curated step-26000 medium sample boxes and provenance

iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/selected-samples/medium-selected-samples-step26000.json

@@ -0,0 +1,41 @@
+{
+  "checkpoint_step": 26000,
+  "generation_protocol": "supplied true level-zero code; random untruncated; not unconditional",
+  "reference_model": "gpt2-large",
+  "reference_dtype": "float32",
+  "tf32_allowed": false,
+  "samples": [
+    {
+      "seed": 0,
+      "index": 70,
+      "text": " but it is usually easy to do how each that can work for it, that would products things that can be done. Andt means there weren’t tools to help how a logic without implementing any specific code.\n\nNow, on what you’ll have with this, it’s easier to try to solve the problem. There are several examples of new language by programming. For now we are some enforce languages, are new languages, and defined information. You can see the code for the language. comment sounds. Many teach toer usingified is to understand language I understand. However, the idea is that I’ve used in the early processes in writing the data, for you are still for such programs. there are also new languages with some parts that are the other language. For example, it would be wrong to one something from one language language. You can write that with three separate parts of your own is under just many a languages.\n\nIt would require a better language language. This means that in a language in an error is considered critical, i.e. If you write a language or better to make a single and; that we can write a language that provides into very different things. I really like to refine my problems",
+      "prompt": "",
+      "token_count": 256,
+      "scored_tokens": 255,
+      "gen_ppl": 76.85733032226562,
+      "entropy_nats": 4.504906832692063,
+      "source_file": "/home/ishank/mstok-results/iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-0.samples.json",
+      "text_sha256": "c5f49753a454c2ab3d8bc7f6284be40204d03b032743eb05497128c6da44ddb6",
+      "replacement_characters": 0,
+      "adjacent_word_repetition": 0.009345794392523364,
+      "repeated_4gram_fraction": 0
+    },
+    {
+      "seed": 4,
+      "index": 55,
+      "text": " back.\n\nWhen control when the weapon breaks, you can even break it on the ground, at it.\n\nYou may take it a a where any move is not a damage, causing an action to the lane.\n\nYou can hit this way with which you want the available allowed. This is cut\n or back of of attack.\n\nWhen you hit one from the opponent to your goal, then your opponent will get further damage or by your attacking attack.\n\nYou can also hit 2 damage from another weapon from another. of another combos.\n\nThe Space.\nThe hit 2 damage your target the damage make your opponent’s move.\n\nThe Clocking\n.Through a Combat attack, catching a player from out to 1, on which a player cannot choose either and the z points will be over there. Again the penalty, a while is at you unless 1 you are near a distance or at a combat mode.\n\nYou may either change your chance ally. When an opponent makes a special break down, it will affect the damage making it and than than the target.\n\nOrder\n\nOnce you only face a target at the same time, you can still face it on the ground facing with your hand.",
+      "prompt": "",
+      "token_count": 256,
+      "scored_tokens": 255,
+      "gen_ppl": 81.66126251220703,
+      "entropy_nats": 4.252252332370423,
+      "source_file": "/home/ishank/mstok-results/iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/random-seed-4.samples.json",
+      "text_sha256": "6581fbbe709a5d3590087d72a9f1446115d697b0db76e9ec178d8c70aa8a5e1d",
+      "replacement_characters": 0,
+      "adjacent_word_repetition": 0.014705882352941176,
+      "repeated_4gram_fraction": 0.004950495049504955
+    }
+  ],
+  "selection_note": "Curated for topic consistency, absence of long repetition loops, and proximity to requested PPL. Both retain conspicuous grammatical and semantic errors; not fluent or representative random examples.",
+  "score_source": "/home/ishank/mstok-results/medium-sample-scores-step26000-20260919.json"
+}

iclr-debug-medium-speed-full-decay-20260918-v1/decoder-mstok/evaluation/step-26000/selected-samples/medium-selected-samples-step26000.tex

@@ -0,0 +1,52 @@
+% Selected, unedited MsTok-medium samples; not random representative examples.
+% Supplied true level-zero code: these are NOT unconditional samples.
+% Individual PPL: GPT-2 large, FP32, TF32 disabled.
+\noindent
+% Checkpoint 26000; seed 0; zero-based sample index 70; 256 tokens.
+\begin{minipage}[t]{0.49\linewidth}
+\begin{tcolorbox}[samplestyle, colback=white, colframe=black]
+\scriptsize\linespread{1.06}\selectfont
+\spheading{black}{Sample 1}{%
+  \textbf{Gen.\ PPL:} 76.86}
+\noindent but it is usually easy to do how each that can work for it, that would products things that can be done. Andt means there weren’t tools to help how a logic without implementing any specific code.
+
+Now, on what you’ll have with this, it’s easier to try to solve the problem. There are several examples of new language by programming. For now we are some enforce languages, are new languages, and defined information. You can see the code for the language. comment sounds. Many teach toer usingified is to understand language I understand. However, the idea is that I’ve used in the early processes in writing the data, for you are still for such programs. there are also new languages with some parts that are the other language. For example, it would be wrong to one something from one language language. You can write that with three separate parts of your own is under just many a languages.
+
+It would require a better language language. This means that in a language in an error is considered critical, i.e. If you write a language or better to make a single and; that we can write a language that provides into very different things. I really like to refine my problems
+\end{tcolorbox}
+\end{minipage}\hfill
+% Checkpoint 26000; seed 4; zero-based sample index 55; 256 tokens.
+\begin{minipage}[t]{0.49\linewidth}
+\begin{tcolorbox}[samplestyle, colback=white, colframe=black]
+\scriptsize\linespread{1.06}\selectfont
+\spheading{black}{Sample 2}{%
+  \textbf{Gen.\ PPL:} 81.66}
+\noindent back.
+
+When control when the weapon breaks, you can even break it on the ground, at it.
+
+You may take it a a where any move is not a damage, causing an action to the lane.
+
+You can hit this way with which you want the available allowed. This is cut
+ or back of of attack.
+
+When you hit one from the opponent to your goal, then your opponent will get further damage or by your attacking attack.
+
+You can also hit 2 damage from another weapon from another. of another combos.
+
+The Space.
+The hit 2 damage your target the damage make your opponent’s move.
+
+The Clocking
+.Through a Combat attack, catching a player from out to 1, on which a player cannot choose either and the z points will be over there. Again the penalty, a while is at you unless 1 you are near a distance or at a combat mode.
+
+You may either change your chance ally. When an opponent makes a special break down, it will affect the damage making it and than than the target.
+
+Order
+
+Once you only face a target at the same time, you can still face it on the ground facing with your hand.
+\end{tcolorbox}
+\end{minipage}
+\par\smallskip
+{\scriptsize Selected 256-token outputs from \name-med.\ (step 26,000), conditioned on the true level-zero code. Generated wording is unchanged.}
+