JANG_6M JANG affine bundle of poolside/Laguna-S-2.1 — 118B-parameter / ~8B-active MoE for agentic coding and long-horizon work, quantized for Apple Silicon.

Attention (incl. the softplus output gate), shared expert, router, and norms stay high precision; the 256-routed-expert bulk carries 6-bit gate/up/down, protected by activation-aware (AWQ) scales chosen by measured quantization error against a no-AWQ baseline (3.1% lower reconstruction error at this bit width).

Bundle

Field	Value
Source	`poolside/Laguna-S-2.1` @ `e80da38`
Architecture	`laguna` — 48 layers (12 global + 36 SWA w512), 118B-A8B, 256 experts top-10 + shared, 1M ctx
On-disk size	96.5 GB (21 shards)
Routed experts	6-bit gate/up/down affine, group 64, AWQ folded
Attention q/k/v/o + g_proj	8-bit affine
Shared expert / dense FFN	8-bit affine
Embeddings / lm_head	6-bit / 8-bit affine
Router, e_score bias, norms	fp16 passthrough
Modality	text-only (verified from tensor index)

Measured (M5 Max, 128 GB, greedy, wired)

Metric	Value
Decode	30.7 tok/s
Long-context cache parity	teacher-forced top-1 agreement 1.000 / 1.000 (pre/post the 512 sliding window, 2,913-token pass)

Chat / reasoning

GLM-style think tags; enable_thinking toggles reasoning (thinking is ON by default in this revision's template; pass enable_thinking=False to disable)
The FULL chat template is inlined in tokenizer_config.json (upstream ships only an {% include %} stub that most runtimes cannot resolve — inlining is what makes the reasoning toggle actually work)
Stop tokens eos_token_id = [2, 24] — id 24 is end-of-turn and must be in the stop set
Template emits its own leading 〈|EOS|〉 (bos 2): do not prepend another
Tool calls: <tool_call>name<arg_key>k</arg_key><arg_value>v</arg_value></tool_call>
Vendor sampling: temperature 1.0, top_p 1.0, top_k 20

Loaders must honor the per-module {bits, group_size, mode} overrides in config.json[quantization].

Quantized and verified by Jinho Jang (eric@jangq.ai). License: OpenMDW-1.1 (inherited).