feat(stage-8c): full-attention layer + decoder + Model + ForCausalLM for qwen3_5

Completes the single-GPU dense path for Qwen3-Next (Qwen3.6's
architecture). The four new modules wrap the substantive
`linear_attn.rs` (landed previously) with the rest of the
transformer:

- `arch/qwen3_5/rope.rs` — text-side rotary embedding. MRoPE is
  simplified to plain RoPE (the three position grids collapse to one
  for text-only inference); uses candle's `rope_slow` for the
  GLM-style rotate-half rotation.
- `arch/qwen3_5/mlp.rs` — Qwen3_5MLP (SwiGLU: gate/up/down, bias=False).
- `arch/qwen3_5/full_attn.rs` — Qwen3_5Attention with the two
  Qwen3-Next quirks:
  - `q_proj` widened to `2 * num_heads * head_dim`; second half
    sigmoid'd and multiplied into the attention output before `o_proj`.
  - q_norm/k_norm use the `(1+w)*x` RmsNorm variant.
- `arch/qwen3_5/decoder.rs` — Qwen3_5DecoderLayer dispatching on
  `layer_types[i]` to either Full attention or GatedDeltaNet.

`arch/qwen3_5/mod.rs` gets the real `Qwen3_5Model` (embedding + layer
stack + final norm) and `Qwen3_5ForCausalLM` (model + lm_head). The
forward returns `[B, 1, vocab]` to match `qwen3_dense`; the harness's
`squeeze_to_vocab` handles either shape.

Switch: `candle.rs::load_arch_dense` for `model_type=qwen3_5` now
builds a `ShardedVarBuilder` instead of a plain VarBuilder. The
sharded backend falls through to the unsharded path when
`world_size=1`, so single-GPU load is zero-cost; this lets the
forthcoming `tp_qwen3_5.rs` reuse the same load functions without a
second copy.

Verified: cargo build CPU + --features cuda inside the patched
container; clippy clean on both; 32 lib tests still pass. The
ForCausalLM forward no longer bails — but numerical correctness vs
the Python reference hasn't been validated yet (that's the next
step, with the Tbilisi probe).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

crates/neuron/src/harness/candle.rs

@@ -617,12 +617,22 @@ impl CandleHarness {
                     })))
                 }
                 "qwen3_5" => {
-                    // Stage 8c scaffold: config parses, model
-                    // constructs, but forward bails. See
-                    // `arch/qwen3_5.rs` for the open architecture work.
+                    // Qwen3-Next needs a ShardedVarBuilder because its
+                    // load functions use the sharded backend (so they
+                    // can be reused unchanged by the future TP variant).
+                    // With world_size=1 the backend falls through to
+                    // the unsharded path, so there is no per-load cost.
                     let cfg: super::arch::qwen3_5::Config = serde_json::from_str(&cfg_text)
                         .context("parse Qwen3-Next (qwen3_5) config.json")?;
-                    let model = super::arch::qwen3_5::Qwen3_5ForCausalLM::new(cfg, vb)
+                    let sharded_vb = unsafe {
+                        candle_nn::var_builder::ShardedSafeTensors::var_builder(
+                            &safetensors_paths,
+                            dtype,
+                            &device_for_load,
+                        )
+                        .context("build ShardedVarBuilder for Qwen3-Next")?
+                    };
+                    let model = super::arch::qwen3_5::Qwen3_5ForCausalLM::new(cfg, sharded_vb)
                         .context("build Qwen3-Next dense model")?;
                     Ok(ModelArch::Qwen3_5Dense(model))
                 }