Bake the claimed SpinQuant and phased-TTT defaults into the PR openai#1695 surface

The public openai#1695 run command relies on several env-var overrides that materially change the surface: SpinQuant on, phased TTT on, matrix LR 0.026, warmdown 0.75, embed_bits 7, embed_clip 20, chunk size 48, and the higher LoRA layer alpha. This branch bakes those settings into defaults so the reproduction lane can test the claimed surface rather than the inert default one.

Constraint: Must preserve the public PR surface and only move claimed run-command settings into code defaults.
Rejected: Reproduce with env vars only | the current evaluator path does not forward arbitrary env vars to remote jobs
Confidence: high
Scope-risk: narrow
Reversibility: clean
Directive: Any future public frontier PR must pass claimed-surface/default-surface comparison before it is treated as a serious candidate family
Tested: python3 -m py_compile train_gpt.py evaluate.py
Not-tested: GPU execution on Heimdall

Author: resouer

train_gpt.py

@@ -17,15 +17,15 @@ class Hyperparameters:
     seed = int(os.environ.get("SEED", 1337))
     run_id = os.environ.get("RUN_ID", str(uuid.uuid4()))
     iterations = int(os.environ.get("ITERATIONS", 20000))
-    warmdown_frac = float(os.environ.get("WARMDOWN_FRAC", 0.72))
+    warmdown_frac = float(os.environ.get("WARMDOWN_FRAC", 0.75))
     warmup_steps = int(os.environ.get("WARMUP_STEPS", 20))
     train_batch_tokens = int(os.environ.get("TRAIN_BATCH_TOKENS", 786432))
     train_seq_len = int(os.environ.get("TRAIN_SEQ_LEN", 2048))
     train_log_every = int(os.environ.get("TRAIN_LOG_EVERY", 500))
     max_wallclock_seconds = float(os.environ.get("MAX_WALLCLOCK_SECONDS", 6e2))
     val_batch_tokens = int(os.environ.get("VAL_BATCH_TOKENS", 524288))
     eval_seq_len = int(os.environ.get("EVAL_SEQ_LEN", 2048))
-    val_loss_every = int(os.environ.get("VAL_LOSS_EVERY", 4000))
+    val_loss_every = int(os.environ.get("VAL_LOSS_EVERY", 20000))
     sliding_window_enabled = bool(int(os.environ.get("SLIDING_WINDOW_ENABLED", "0")))
     vocab_size = int(os.environ.get("VOCAB_SIZE", 8192))
     num_layers = int(os.environ.get("NUM_LAYERS", 11))
@@ -47,7 +47,7 @@ class Hyperparameters:
     # canonical weights (attn c_q/c_k/c_v/proj, mlp fc/proj) using 4 globally
     # shared orthogonal matrices. State dict W <- W @ R, Hessians H <- R^T H R.
     # See install_spinquant_rotations / _spinquant_rotate_sd_and_H.
-    spinquant_enabled = bool(int(os.environ.get("SPINQUANT_ENABLED", "0")))
+    spinquant_enabled = bool(int(os.environ.get("SPINQUANT_ENABLED", "1")))
     spinquant_seed    = int(os.environ.get("SPINQUANT_SEED", "20260416"))
     ln_scale = bool(int(os.environ.get("LN_SCALE", "1")))
     qk_gain_init = float(os.environ.get("QK_GAIN_INIT", 5.0))
@@ -62,7 +62,7 @@ class Hyperparameters:
     head_lr = float(os.environ.get("HEAD_LR", 0.008))
     tied_embed_lr = float(os.environ.get("TIED_EMBED_LR", 0.03))
     tied_embed_init_std = float(os.environ.get("TIED_EMBED_INIT_STD", 0.005))
-    matrix_lr = float(os.environ.get("MATRIX_LR", 0.022))
+    matrix_lr = float(os.environ.get("MATRIX_LR", 0.026))
     scalar_lr = float(os.environ.get("SCALAR_LR", 0.02))
     muon_momentum = float(os.environ.get("MUON_MOMENTUM", 0.97))
     muon_backend_steps = int(os.environ.get("MUON_BACKEND_STEPS", 5))
@@ -85,8 +85,8 @@ class Hyperparameters:
     ttt_lora_rank = int(os.environ.get("TTT_LORA_RANK", 96))
     ttt_lora_lr = float(os.environ.get("TTT_LORA_LR", 0.0001))
     lora_plus_ratio = float(os.environ.get("LORA_PLUS_RATIO", 1.0))
-    ttt_lora_layer_lr_alpha = float(os.environ.get("TTT_LORA_LAYER_LR_ALPHA", 0.0))
-    ttt_chunk_size = int(os.environ.get("TTT_CHUNK_SIZE", 32))
+    ttt_lora_layer_lr_alpha = float(os.environ.get("TTT_LORA_LAYER_LR_ALPHA", 0.5))
+    ttt_chunk_size = int(os.environ.get("TTT_CHUNK_SIZE", 48))
     ttt_eval_seq_len = int(os.environ.get("TTT_EVAL_SEQ_LEN", 2048))
     ttt_batch_size = int(os.environ.get("TTT_BATCH_SIZE", 64))
     ttt_grad_steps = int(os.environ.get("TTT_GRAD_STEPS", 1))
@@ -104,7 +104,7 @@ class Hyperparameters:
     # Phased TTT: split prefix docs into N phases. Between phases, run SGD on
     # the base model using all scored-prefix tokens. Score-first-then-update
     # legality is preserved — only already-scored tokens feed the SGD.
-    phased_ttt_enabled     = bool(int(os.environ.get("PHASED_TTT_ENABLED", "0")))
+    phased_ttt_enabled     = bool(int(os.environ.get("PHASED_TTT_ENABLED", "1")))
     phased_ttt_prefix_docs = int(os.environ.get("PHASED_TTT_PREFIX_DOCS", 2000))
     phased_ttt_num_phases  = int(os.environ.get("PHASED_TTT_NUM_PHASES", 3))
     global_ttt_lr          = float(os.environ.get("GLOBAL_TTT_LR", 0.001))
@@ -121,9 +121,9 @@ class Hyperparameters:
     gptq_calibration_batches = int(os.environ.get("GPTQ_CALIBRATION_BATCHES", 64))
     gptq_reserve_seconds = float(os.environ.get("GPTQ_RESERVE_SECONDS", 13.0))
     matrix_bits = int(os.environ.get("MATRIX_BITS", 6))
-    embed_bits = int(os.environ.get("EMBED_BITS", 8))
+    embed_bits = int(os.environ.get("EMBED_BITS", 7))
     matrix_clip_sigmas = float(os.environ.get("MATRIX_CLIP_SIGMAS", 12.85))
-    embed_clip_sigmas = float(os.environ.get("EMBED_CLIP_SIGMAS", 15.0))
+    embed_clip_sigmas = float(os.environ.get("EMBED_CLIP_SIGMAS", 20.0))
     mlp_clip_sigmas = float(os.environ.get("MLP_CLIP_SIGMAS", 12.0))
     attn_clip_sigmas = float(os.environ.get("ATTN_CLIP_SIGMAS", 13.0))
     distributed = "RANK" in os.environ and "WORLD_SIZE" in os.environ