Initial private pilot 3 adapter upload

.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 +1,220 @@
+---
+language:
+  - en
+license: mit
+base_model:
+  - Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-v2
+library_name: transformers
+pipeline_tag: text-generation
+tags:
+  - reasoning
+  - evidence-grounding
+  - attribution
+  - fever
+  - hotpotqa
+  - lora
+  - peft
+  - distillation
+  - research
+---
+
+# Evidence-Faithful Reasoning Pilot 3
+
+This release is the standalone-model-v2 checkpoint from the
+evidence-faithful reasoning project.
+
+The project studies a stricter target than ordinary question answering. The
+system must answer from a bounded packet of source text, identify the right
+evidence, quote exact supporting text, and abstain with
+`Insufficient evidence.` when the packet does not justify a claim.
+
+Pilot 3 is the strongest standalone model artifact produced by the project. It
+is the model this release should lead with for direct download and use.
+
+## What This Release Is
+
+This card is for the pilot 3 standalone adapter release:
+
+- checkpoint:
+  `outputs/sft_v1_v2_teacher_distill_pilot_v3_partialdev/checkpoint-16`
+
+Pilot 3 packages the strongest version of the project’s evidence-faithful
+behavior that successfully moved into one model. The same project also
+produced a benchmark-winning hybrid system for the frozen held-out benchmark,
+but this page is for the main downloadable model artifact.
+
+## Why It Matters
+
+Pilot 3 is important for three reasons:
+
+1. it is the first standalone model in the project that stays strong across
+   multiple non-`probe_v0` evaluation surfaces
+2. it materially improves raw quote-faithful behavior over the earlier bridge
+   model
+3. it gives the project a clean downloadable model artifact, not only a
+   benchmark-facing system result
+
+## At A Glance
+
+- model role:
+  main downloadable model from the project
+- benchmark-facing winner:
+  bridge checkpoint-2 + `deterministic_v3`
+- public model story:
+  pilot 3 is the strongest direct model artifact; the hybrid stack remains the
+  strongest full benchmark system
+
+## Quick Start
+
+Pilot 3 is a LoRA adapter on top of the base model listed above. To use it as
+the direct standalone release, load the base model and attach the published
+pilot 3 adapter:
+
+```python
+from peft import PeftModel
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+base_id = "Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-v2"
+adapter_id = "darcar0/evidence-faithful-reasoning-pilot-3"
+
+tokenizer = AutoTokenizer.from_pretrained(base_id)
+base = AutoModelForCausalLM.from_pretrained(base_id, device_map="auto")
+model = PeftModel.from_pretrained(base, adapter_id)
+```
+
+If you want to reproduce the benchmark-winning hybrid result instead, use the
+project repository’s bridge checkpoint plus `deterministic_v3`.
+
+## Strongest Public Evidence
+
+Pilot 3 holds up across all three non-`probe_v0` surfaces used for standalone
+selection:
+
+### Fixed dev triage slice
+
+Pilot 3 + `deterministic_v3`:
+
+- task `1.0000`
+- strict `0.6190`
+- evidence F1 `0.8320`
+- quote F1 `0.7095`
+
+### Untouched 104-task Hotpot shadow slice
+
+- pilot 3 raw materially improved quote-faithful behavior over raw bridge
+- pilot 3 + `deterministic_v3` matched bridge + `deterministic_v3`
+
+### Fresh 36-task mixed public holdout
+
+Bridge + `deterministic_v3`:
+
+- task `0.8611`
+- strict `0.5833`
+- evidence F1 `0.8815`
+- quote F1 `0.8815`
+
+Pilot 3 + `deterministic_v3`:
+
+- task `0.8889`
+- strict `0.5833`
+- evidence F1 `0.9093`
+- quote F1 `0.9093`
+
+Raw bridge vs raw pilot 3 on the same holdout:
+
+- bridge raw
+  - task `0.8611`
+  - strict `0.2222`
+  - evidence F1 `0.8815`
+  - quote F1 `0.3343`
+- pilot 3 raw
+  - task `0.8889`
+  - strict `0.4444`
+  - evidence F1 `0.9093`
+  - quote F1 `0.6815`
+
+This means pilot 3 is both the cleanest release artifact from the project and
+a materially stronger standalone model than the earlier bridge checkpoint on a
+fresh public holdout.
+
+## How This Fits Into The Full Project
+
+The project has two finished outcomes:
+
+1. **Benchmark-facing winner**
+   - bridge checkpoint-2 + `deterministic_v3`
+   - perfect on frozen held-out `probe_v0`
+2. **Main downloadable model**
+   - pilot 3
+   - strongest standalone model released from the project
+
+That distinction is deliberate.
+
+The hybrid stack is the strongest full benchmark system. Pilot 3 is the
+strongest clean model artifact.
+
+## Intended Use
+
+This release is intended for:
+
+- research on evidence-faithful reasoning
+- bounded document QA with explicit evidence requirements
+- claim verification and grounded QA from fixed evidence packets
+- policy, compliance, contract, or internal-document reasoning workflows where
+  answers must be justified from a fixed packet of text
+
+This is a specialized grounded reasoning model, not a general-purpose chatbot
+replacement.
+
+## Training / Data Notes
+
+The training and evaluation surfaces are public-data-backed and derived from:
+
+- FEVER-style verify-claim data
+- HotpotQA-style grounded QA data
+- project-local bounded packet scaffolds
+
+The held-out benchmark `probe_v0` remained frozen and was **not** used as a
+tuning surface for this standalone selection cycle.
+
+## Why Pilot 3 Was Frozen
+
+Pilot 4 was a deliberately narrow refinement meant to fix one FEVER
+month/date-style temporal insufficiency error.
+
+It fixed that targeted row.
+
+But it also weakened broader behavior on the larger evaluation surfaces.
+
+That made pilot 4 a stop signal, not a better release.
+
+So pilot 3 is the checkpoint the project froze as the strongest standalone
+model release.
+
+## Limitations
+
+- Pilot 3 is the main downloadable model from the project. The separate
+  benchmark-facing winner is bridge checkpoint-2 + `deterministic_v3`.
+- Perfect `probe_v0` belongs to the hybrid stack, not to pilot 3 alone.
+- Perfect `probe_v0` should not be interpreted as proof of general faithful
+  reasoning.
+- This release is specialized around bounded evidence packets, not general
+  chat.
+
+## Repository Surfaces
+
+Canonical project surfaces:
+
+- standalone freeze memo:
+  `reports/standalone_model_v2_freeze_memo.md`
+- fresh standalone holdout comparison:
+  `reports/standalone_model_v2_holdout_v1_bridge_vs_pilot3_status.md`
+- benchmark-facing final artifact:
+  `reports/sft_v1_final_artifact_status.md`
+- release page:
+  `release/index.html`
+
+## Citation
+
+If you use this release, cite the project repository and the standalone freeze
+memo.

adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-v2",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.1,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "o_proj",
+    "k_proj",
+    "down_proj",
+    "gate_proj",
+    "v_proj",
+    "up_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57479e3f6df4b0838dcb5615ba8e10253cdbd4cecdfe193b8bf0fca3caa85d04
+size 159452256

chat_template.jinja

@@ -0,0 +1,154 @@
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- macro render_content(content, do_vision_count, is_system_content=false) %}
+    {%- if content is string %}
+        {{- content }}
+    {%- elif content is iterable and content is not mapping %}
+        {%- for item in content %}
+            {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain images.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Picture ' ~ image_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|image_pad|><|vision_end|>' }}
+            {%- elif 'video' in item or item.type == 'video' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain videos.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Video ' ~ video_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|video_pad|><|vision_end|>' }}
+            {%- elif 'text' in item %}
+                {{- item.text }}
+            {%- else %}
+                {{- raise_exception('Unexpected item type in content.') }}
+            {%- endif %}
+        {%- endfor %}
+    {%- elif content is none or content is undefined %}
+        {{- '' }}
+    {%- else %}
+        {{- raise_exception('Unexpected content type.') }}
+    {%- endif %}
+{%- endmacro %}
+{%- if not messages %}
+    {{- raise_exception('No messages provided.') }}
+{%- endif %}
+{%- if tools and tools is iterable and tools is not mapping %}
+    {{- '<|im_start|>system\n' }}
+    {{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>" }}
+    {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {%- if content %}
+            {{- '\n\n' + content }}
+        {%- endif %}
+    {%- endif %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {%- set content = render_content(messages[0].content, false, true)|trim %}
+        {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if ns.multi_step_tool and message.role == "user" %}
+        {%- set content = render_content(message.content, false)|trim %}
+        {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
+            {%- set ns.multi_step_tool = false %}
+            {%- set ns.last_query_index = index %}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if ns.multi_step_tool %}
+    {{- raise_exception('No user query found in messages.') }}
+{%- endif %}
+{%- for message in messages %}
+    {%- set content = render_content(message.content, true)|trim %}
+    {%- if message.role == "system" %}
+        {%- if not loop.first %}
+            {{- raise_exception('System message must be at the beginning.') }}
+        {%- endif %}
+    {%- elif message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is string %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in content %}
+                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- set reasoning_content = reasoning_content|trim %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if tool_call.function is defined %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {%- if loop.first %}
+                    {%- if content|trim %}
+                        {{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- else %}
+                        {{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                    {%- endif %}
+                {%- else %}
+                    {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+                {%- endif %}
+                {%- if tool_call.arguments is defined %}
+                    {%- for args_name, args_value in tool_call.arguments|items %}
+                        {{- '<parameter=' + args_name + '>\n' }}
+                        {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
+                        {{- args_value }}
+                        {{- '\n</parameter>\n' }}
+                    {%- endfor %}
+                {%- endif %}
+                {{- '</function>\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.previtem and loop.previtem.role != "tool" %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if not loop.last and loop.nextitem.role != "tool" %}
+            {{- '<|im_end|>\n' }}
+        {%- elif loop.last %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- else %}
+        {{- raise_exception('Unexpected message role.') }}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- else %}
+        {{- '<think>\n' }}
+    {%- endif %}
+{%- endif %}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4
+size 19989343

tokenizer_config.json

@@ -0,0 +1,31 @@
+{
+  "add_prefix_space": false,
+  "audio_bos_token": "<|audio_start|>",
+  "audio_eos_token": "<|audio_end|>",
+  "audio_token": "<|audio_pad|>",
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "image_token": "<|image_pad|>",
+  "is_local": false,
+  "model_max_length": 262144,
+  "model_specific_special_tokens": {
+    "audio_bos_token": "<|audio_start|>",
+    "audio_eos_token": "<|audio_end|>",
+    "audio_token": "<|audio_pad|>",
+    "image_token": "<|image_pad|>",
+    "video_token": "<|video_pad|>",
+    "vision_bos_token": "<|vision_start|>",
+    "vision_eos_token": "<|vision_end|>"
+  },
+  "pad_token": "<|endoftext|>",
+  "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
+  "split_special_tokens": false,
+  "tokenizer_class": "TokenizersBackend",
+  "unk_token": null,
+  "video_token": "<|video_pad|>",
+  "vision_bos_token": "<|vision_start|>",
+  "vision_eos_token": "<|vision_end|>"
+}