Initial release: R14 endpoint (sibling-distilled capability-floor probe)

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,107 @@
+---
+license: gemma
+language:
+- en
+- kk
+base_model:
+- google/gemma-3-270m
+pipeline_tag: text-generation
+tags:
+- edge-cloud-routing
+- verbalized-confidence
+- self-aware
+- routing
+- continual-learning
+- multi-round
+- sibling-distilled
+- capability-floor
+library_name: transformers
+---
+
+# FogGen (Gemma-3-270m, sibling-distilled): capability-floor R14 endpoint
+
+> **The 270M-parameter capability-floor probe of the FogGen recipe.** Sibling-distilled from the Gemma-3-1b-it buffer to install the FogGen output format, then run through the same 14-round self-evolving chain. Demonstrates the recipe pays off at deployment-grade magnitudes from roughly 0.6B upward; below that, lift becomes order-of-magnitude smaller and a sibling-distilled SFT pass is required to install the format at all.
+
+This is a **capability-floor diagnostic checkpoint**, not a deployment model. The canonical deployment endpoint is [`issai/foggen`](https://huggingface.co/issai/foggen) at the 0.6B scale.
+
+For background on the system overview, training pipeline, and routing protocol, see the [`issai/foggen`](https://huggingface.co/issai/foggen) model card.
+
+## Why this exists
+
+Native zero-shot routing is infeasible at the 270M scale: no prompting or constrained-decoding setup we tried exceeded 54% format compliance on the FogGen output schema (the model fails to emit the `Confidence:`/`Final answer:` pattern reliably enough to extract a routing signal). We therefore probe this scale with a two-stage protocol:
+
+1. **Sibling-distillation SFT pass** — one round of SFT on the calibration buffer of the [Gemma-3-1b-it sibling](https://huggingface.co/issai/foggen-gemma3-1b), using the larger model's bucket labels as targets. This installs the FogGen format on the 270M backbone.
+2. **Standard 14-round chain** — identical recipe to [`issai/foggen`](https://huggingface.co/issai/foggen) from there: 7 domain rotation, LoRA r=16 α=32, bf16, 2 epochs/round, same cloud teacher.
+
+The released checkpoint is R14 of the post-distillation chain.
+
+## Performance
+
+System accuracy at τ=0.5 on the seven MCQ domains (full test sets, ~16,200 queries). Cloud baseline is Qwen3-30B-A3B-Instruct-2507.
+
+| Domain | Cloud only | R14 raw | Random @ τ=0.5 | **FogGen @ τ=0.5** | Cloud routed |
+|---|---|---|---|---|---|
+| Finance | 69.5% | 32.2% | 58.2% | **60.2%** | 69.5% |
+| Science | 72.7% | 30.4% | 58.2% | **59.5%** | 65.6% |
+| Coding | 74.2% | 34.3% | 64.7% | **65.7%** | 76.3% |
+| Law | 70.7% | 31.7% | 58.5% | **59.7%** | 68.7% |
+| Math | 60.1% | 24.5% | 58.3% | **58.5%** | 94.9% |
+| Kazakh culture | 95.8% | 43.7% | 60.3% | **59.3%** | 31.9% |
+| Medical | 74.0% | 32.2% | 59.8% | **60.8%** | 65.9% |
+| **Mean** | **73.9%** | **32.7%** | **59.7%** | **60.5%** | **67.5%** |
+
+**Routing benefit over Random: +0.8 percentage points mean at τ=0.5** — positive on six of seven domains, negative on Kazakh culture (the headroom-collapse domain).
+
+Compared to [`issai/foggen`](https://huggingface.co/issai/foggen) (+4.6 pp at 0.6B) and [`issai/foggen-gemma3-1b`](https://huggingface.co/issai/foggen-gemma3-1b) (+5.9 pp at 1B), the lift here is an order of magnitude smaller. The recipe still produces positive lift, but the magnitude scales sharply with edge capacity below the 0.6B mark.
+
+## When to use this
+
+This checkpoint is intended for:
+
+- **Reproducing the capability-floor analysis** in the paper.
+- **Studying the sibling-distillation bootstrap** as a technique for installing a calibrated output format on edge models too small to host it natively.
+- **Tightly memory-constrained deployments** where 270M is the only feasible size and a modest routing lift over Random is still preferable to no routing at all.
+
+For any deployment with a 0.6B-or-larger budget, prefer [`issai/foggen`](https://huggingface.co/issai/foggen).
+
+## Quick demo
+
+```python
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+model = AutoModelForCausalLM.from_pretrained("issai/foggen-gemma3-270m", torch_dtype="bfloat16", device_map="auto")
+tokenizer = AutoTokenizer.from_pretrained("issai/foggen-gemma3-270m")
+
+SYSTEM = """You are a self-aware multiple-choice assistant.
+
+Rules:
+- First, assess your confidence in solving this question.
+- Then give your answer.
+- Output format:
+  Confidence: <0.0|0.25|0.5|0.75|1.0>
+  Final answer: <OPTION_LETTER>"""
+
+messages = [
+    {"role": "system", "content": SYSTEM},
+    {"role": "user", "content": "<your MCQ here>"},
+]
+inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True).to(model.device)
+outputs = model.generate(inputs, max_new_tokens=64, do_sample=False)
+print(tokenizer.decode(outputs[0][inputs.shape[-1]:], skip_special_tokens=True))
+```
+
+The routing decision (`route_query` helper, threshold τ) is identical to the [`issai/foggen`](https://huggingface.co/issai/foggen) card.
+
+## License
+
+Inherits the [Gemma Terms of Use](https://ai.google.dev/gemma/terms) from google/gemma-3-270m.
+
+## Limitations
+
+- **Format compliance is bootstrap-dependent.** Without the sibling-distillation pass the model cannot reliably emit the FogGen output schema; the released checkpoint inherits the format installation from its 1B sibling.
+- **Small absolute lift.** +0.8 pp mean over Random — the recipe pays off here but at a magnitude an order smaller than at the 0.6B–1.7B scales tested in the paper.
+- **Kazakh culture regression.** −1.0 pp at τ=0.5 — the headroom-collapse domain is also the one place the recipe slips below Random at this scale.
+
+## Citation
+
+See [`issai/foggen`](https://huggingface.co/issai/foggen) for the paper citation.

added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<image_soft_token>": 262144
+}

chat_template.jinja

@@ -0,0 +1,47 @@
+{{ bos_token }}
+{%- if messages[0]['role'] == 'system' -%}
+    {%- if messages[0]['content'] is string -%}
+        {%- set first_user_prefix = messages[0]['content'] + '
+
+' -%}
+    {%- else -%}
+        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '
+
+' -%}
+    {%- endif -%}
+    {%- set loop_messages = messages[1:] -%}
+{%- else -%}
+    {%- set first_user_prefix = "" -%}
+    {%- set loop_messages = messages -%}
+{%- endif -%}
+{%- for message in loop_messages -%}
+    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
+        {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
+    {%- endif -%}
+    {%- if (message['role'] == 'assistant') -%}
+        {%- set role = "model" -%}
+    {%- else -%}
+        {%- set role = message['role'] -%}
+    {%- endif -%}
+    {{ '<start_of_turn>' + role + '
+' + (first_user_prefix if loop.first else "") }}
+    {%- if message['content'] is string -%}
+        {{ message['content'] | trim }}
+    {%- elif message['content'] is iterable -%}
+        {%- for item in message['content'] -%}
+            {%- if item['type'] == 'image' -%}
+                {{ '<start_of_image>' }}
+            {%- elif item['type'] == 'text' -%}
+                {{ item['text'] | trim }}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- else -%}
+        {{ raise_exception("Invalid content type") }}
+    {%- endif -%}
+    {{ '<end_of_turn>
+' }}
+{%- endfor -%}
+{%- if add_generation_prompt -%}
+    {{'<start_of_turn>model
+'}}
+{%- endif -%}

config.json

@@ -0,0 +1,54 @@
+{
+  "_sliding_window_pattern": 6,
+  "architectures": [
+    "Gemma3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "attn_logit_softcapping": null,
+  "bos_token_id": 2,
+  "dtype": "bfloat16",
+  "eos_token_id": 1,
+  "final_logit_softcapping": null,
+  "head_dim": 256,
+  "hidden_activation": "gelu_pytorch_tanh",
+  "hidden_size": 640,
+  "initializer_range": 0.02,
+  "intermediate_size": 2048,
+  "layer_types": [
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "model_type": "gemma3_text",
+  "num_attention_heads": 4,
+  "num_hidden_layers": 18,
+  "num_key_value_heads": 1,
+  "pad_token_id": 0,
+  "query_pre_attn_scalar": 256,
+  "rms_norm_eps": 1e-06,
+  "rope_local_base_freq": 10000.0,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 512,
+  "transformers_version": "4.57.6",
+  "use_bidirectional_attention": false,
+  "use_cache": true,
+  "vocab_size": 262144
+}

generation_config.json

@@ -0,0 +1,11 @@
+{
+  "cache_implementation": "hybrid",
+  "do_sample": true,
+  "eos_token_id": [
+    1,
+    106
+  ],
+  "top_k": 64,
+  "top_p": 0.95,
+  "transformers_version": "4.57.6"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4596b28b96c4ccd194b213b9be40ef896909625441a3757718854e21239362c0
+size 536223056

special_tokens_map.json

@@ -0,0 +1,33 @@
+{
+  "boi_token": "<start_of_image>",
+  "bos_token": {
+    "content": "<bos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eoi_token": "<end_of_image>",
+  "eos_token": {
+    "content": "<eos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "image_token": "<image_soft_token>",
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4667f2089529e8e7657cfb6d1c19910ae71ff5f28aa7ab2ff2763330affad795
+size 33384568

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
+size 4689074