Add files using upload-large-folder tool

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test1/README.md

@@ -0,0 +1,58 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: transformers
+model_name: Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test1
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test1
+
+This model is a fine-tuned version of [Qwen/Qwen3-14B-Base](https://huggingface.co/Qwen/Qwen3-14B-Base).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="None", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/katriin-kukk/Cross_lingual_morphological_generalization/runs/r2wxxlcd) 
+
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.29.0
+- Transformers: 5.5.4
+- Pytorch: 2.10.0
+- Datasets: 4.6.1
+- Tokenizers: 0.22.2
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@software{vonwerra2020trl,
+  title   = {{TRL: Transformers Reinforcement Learning}},
+  author  = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
+  license = {Apache-2.0},
+  url     = {https://github.com/huggingface/trl},
+  year    = {2020}
+}
+```

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/README.md

@@ -0,0 +1,58 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: transformers
+model_name: Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2
+
+This model is a fine-tuned version of [Qwen/Qwen3-14B-Base](https://huggingface.co/Qwen/Qwen3-14B-Base).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="None", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/katriin-kukk/Cross_lingual_morphological_generalization/runs/daxtch08) 
+
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.29.0
+- Transformers: 5.5.4
+- Pytorch: 2.10.0
+- Datasets: 4.6.1
+- Tokenizers: 0.22.2
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@software{vonwerra2020trl,
+  title   = {{TRL: Transformers Reinforcement Learning}},
+  author  = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
+  license = {Apache-2.0},
+  url     = {https://github.com/huggingface/trl},
+  year    = {2020}
+}
+```

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1224/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1224/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 256,
+  "lora_bias": false,
+  "lora_dropout": 0.08935122001117755,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 128,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "o_proj",
+    "k_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj",
+    "down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1224/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1224/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1224/trainer_state.json

@@ -0,0 +1,307 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1224,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.4665306943655014,
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.8344842791557312,
+      "learning_rate": 2.1010394822962346e-05,
+      "loss": 1.379196014404297,
+      "mean_token_accuracy": 0.6848452094197274,
+      "num_tokens": 137019.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.7834547007083893,
+      "epoch": 0.24554941682013506,
+      "grad_norm": 1.1248278617858887,
+      "learning_rate": 4.2449573213740254e-05,
+      "loss": 0.7345225524902343,
+      "mean_token_accuracy": 0.7882961884140969,
+      "num_tokens": 267448.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.7194462446868419,
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.6498261094093323,
+      "learning_rate": 6.388875160451815e-05,
+      "loss": 0.664264907836914,
+      "mean_token_accuracy": 0.8042438223958015,
+      "num_tokens": 409071.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.6771699503064156,
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.6653317809104919,
+      "learning_rate": 8.532792999529606e-05,
+      "loss": 0.6322076797485352,
+      "mean_token_accuracy": 0.8130025643110276,
+      "num_tokens": 542207.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.6632554066181183,
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.44293469190597534,
+      "learning_rate": 0.00010676710838607398,
+      "loss": 0.617110595703125,
+      "mean_token_accuracy": 0.8168379080295562,
+      "num_tokens": 679619.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.6508415986597538,
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.49295490980148315,
+      "learning_rate": 0.00012820628677685187,
+      "loss": 0.6023779296875,
+      "mean_token_accuracy": 0.8187075120210647,
+      "num_tokens": 815397.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.6380941693484783,
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.5855488181114197,
+      "learning_rate": 0.00014964546516762976,
+      "loss": 0.5912636947631836,
+      "mean_token_accuracy": 0.8227647486329078,
+      "num_tokens": 949315.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.6346297286450863,
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.3817698657512665,
+      "learning_rate": 0.00017108464355840766,
+      "loss": 0.5902091598510742,
+      "mean_token_accuracy": 0.8231476756930352,
+      "num_tokens": 1082903.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6600051839011056,
+      "eval_loss": 0.5857571363449097,
+      "eval_mean_token_accuracy": 0.8248803649629866,
+      "eval_num_tokens": 1101768.0,
+      "eval_runtime": 106.3263,
+      "eval_samples_per_second": 13.12,
+      "eval_steps_per_second": 1.646,
+      "step": 408
+    },
+    {
+      "entropy": 0.5965584144374441,
+      "epoch": 1.1031307550644567,
+      "grad_norm": 0.5178472399711609,
+      "learning_rate": 0.00017488988654901867,
+      "loss": 0.5578182220458985,
+      "mean_token_accuracy": 0.830260409921559,
+      "num_tokens": 1223073.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.5831904768943786,
+      "epoch": 1.2259054634745243,
+      "grad_norm": 0.4383259117603302,
+      "learning_rate": 0.00017467872603358825,
+      "loss": 0.5399176788330078,
+      "mean_token_accuracy": 0.8351450267434121,
+      "num_tokens": 1360978.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.5885515750944614,
+      "epoch": 1.3486801718845918,
+      "grad_norm": 0.5418892502784729,
+      "learning_rate": 0.00017430800739444695,
+      "loss": 0.5402648544311524,
+      "mean_token_accuracy": 0.8344706323742866,
+      "num_tokens": 1492506.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.603163862079382,
+      "epoch": 1.4714548802946594,
+      "grad_norm": 0.5235878825187683,
+      "learning_rate": 0.0001737784089172037,
+      "loss": 0.5534031677246094,
+      "mean_token_accuracy": 0.8304652521014213,
+      "num_tokens": 1622275.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.5745155589282512,
+      "epoch": 1.5942295887047269,
+      "grad_norm": 0.46109989285469055,
+      "learning_rate": 0.00017309089958207074,
+      "loss": 0.5251047134399414,
+      "mean_token_accuracy": 0.8380702176690101,
+      "num_tokens": 1756450.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.5756107035279274,
+      "epoch": 1.7170042971147943,
+      "grad_norm": 0.38209712505340576,
+      "learning_rate": 0.00017224673729096835,
+      "loss": 0.5253059387207031,
+      "mean_token_accuracy": 0.8379077833890914,
+      "num_tokens": 1890634.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.5695902299880982,
+      "epoch": 1.839779005524862,
+      "grad_norm": 0.4063270688056946,
+      "learning_rate": 0.00017124746656600374,
+      "loss": 0.5199185180664062,
+      "mean_token_accuracy": 0.8399301299452782,
+      "num_tokens": 2028046.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.5635769453644752,
+      "epoch": 1.9625537139349294,
+      "grad_norm": 0.4165436029434204,
+      "learning_rate": 0.00017009491572353503,
+      "loss": 0.5156089019775391,
+      "mean_token_accuracy": 0.8404546654224396,
+      "num_tokens": 2163703.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.5587481154714312,
+      "eval_loss": 0.5402795672416687,
+      "eval_mean_token_accuracy": 0.8386626076698304,
+      "eval_num_tokens": 2203536.0,
+      "eval_runtime": 105.9883,
+      "eval_samples_per_second": 13.162,
+      "eval_steps_per_second": 1.651,
+      "step": 816
+    },
+    {
+      "entropy": 0.5002549305785126,
+      "epoch": 2.0834868017188457,
+      "grad_norm": 0.4061429500579834,
+      "learning_rate": 0.00016879119352899107,
+      "loss": 0.4499735641479492,
+      "mean_token_accuracy": 0.8566478448470837,
+      "num_tokens": 2298231.0,
+      "step": 850
+    },
+    {
+      "entropy": 0.4894501394033432,
+      "epoch": 2.2062615101289134,
+      "grad_norm": 0.530322790145874,
+      "learning_rate": 0.00016733868533856737,
+      "loss": 0.4376458740234375,
+      "mean_token_accuracy": 0.859256779551506,
+      "num_tokens": 2427608.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.48188683822751044,
+      "epoch": 2.329036218538981,
+      "grad_norm": 0.4338841438293457,
+      "learning_rate": 0.0001657400487348575,
+      "loss": 0.435361328125,
+      "mean_token_accuracy": 0.8610378202795982,
+      "num_tokens": 2565982.0,
+      "step": 950
+    },
+    {
+      "entropy": 0.4992289599776268,
+      "epoch": 2.4518109269490487,
+      "grad_norm": 0.44097110629081726,
+      "learning_rate": 0.00016399820866440517,
+      "loss": 0.4474017333984375,
+      "mean_token_accuracy": 0.8565557089447975,
+      "num_tokens": 2695545.0,
+      "step": 1000
+    },
+    {
+      "entropy": 0.49334951683878897,
+      "epoch": 2.574585635359116,
+      "grad_norm": 0.48084133863449097,
+      "learning_rate": 0.00016211635208607405,
+      "loss": 0.440751953125,
+      "mean_token_accuracy": 0.8580261027812958,
+      "num_tokens": 2829781.0,
+      "step": 1050
+    },
+    {
+      "entropy": 0.4898635870218277,
+      "epoch": 2.6973603437691835,
+      "grad_norm": 0.40215447545051575,
+      "learning_rate": 0.00016009792214002617,
+      "loss": 0.44588775634765626,
+      "mean_token_accuracy": 0.8572473660111427,
+      "num_tokens": 2970162.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.4932038363814354,
+      "epoch": 2.820135052179251,
+      "grad_norm": 0.4465612769126892,
+      "learning_rate": 0.00015794661184797816,
+      "loss": 0.4444055938720703,
+      "mean_token_accuracy": 0.8583759877085686,
+      "num_tokens": 3103775.0,
+      "step": 1150
+    },
+    {
+      "entropy": 0.4850301502645016,
+      "epoch": 2.942909760589319,
+      "grad_norm": 0.43597331643104553,
+      "learning_rate": 0.0001556663573562617,
+      "loss": 0.4382463073730469,
+      "mean_token_accuracy": 0.8603282216191291,
+      "num_tokens": 3241634.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.4902002043383462,
+      "eval_loss": 0.5276822447776794,
+      "eval_mean_token_accuracy": 0.8454030500139509,
+      "eval_num_tokens": 3305304.0,
+      "eval_runtime": 106.004,
+      "eval_samples_per_second": 13.16,
+      "eval_steps_per_second": 1.651,
+      "step": 1224
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4080,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.5230462593101824e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1632/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1632/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 256,
+  "lora_bias": false,
+  "lora_dropout": 0.08935122001117755,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 128,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "o_proj",
+    "k_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj",
+    "down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1632/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1632/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-1632/trainer_state.json

@@ -0,0 +1,398 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1632,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.4665306943655014,
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.8344842791557312,
+      "learning_rate": 2.1010394822962346e-05,
+      "loss": 1.379196014404297,
+      "mean_token_accuracy": 0.6848452094197274,
+      "num_tokens": 137019.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.7834547007083893,
+      "epoch": 0.24554941682013506,
+      "grad_norm": 1.1248278617858887,
+      "learning_rate": 4.2449573213740254e-05,
+      "loss": 0.7345225524902343,
+      "mean_token_accuracy": 0.7882961884140969,
+      "num_tokens": 267448.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.7194462446868419,
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.6498261094093323,
+      "learning_rate": 6.388875160451815e-05,
+      "loss": 0.664264907836914,
+      "mean_token_accuracy": 0.8042438223958015,
+      "num_tokens": 409071.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.6771699503064156,
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.6653317809104919,
+      "learning_rate": 8.532792999529606e-05,
+      "loss": 0.6322076797485352,
+      "mean_token_accuracy": 0.8130025643110276,
+      "num_tokens": 542207.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.6632554066181183,
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.44293469190597534,
+      "learning_rate": 0.00010676710838607398,
+      "loss": 0.617110595703125,
+      "mean_token_accuracy": 0.8168379080295562,
+      "num_tokens": 679619.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.6508415986597538,
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.49295490980148315,
+      "learning_rate": 0.00012820628677685187,
+      "loss": 0.6023779296875,
+      "mean_token_accuracy": 0.8187075120210647,
+      "num_tokens": 815397.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.6380941693484783,
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.5855488181114197,
+      "learning_rate": 0.00014964546516762976,
+      "loss": 0.5912636947631836,
+      "mean_token_accuracy": 0.8227647486329078,
+      "num_tokens": 949315.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.6346297286450863,
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.3817698657512665,
+      "learning_rate": 0.00017108464355840766,
+      "loss": 0.5902091598510742,
+      "mean_token_accuracy": 0.8231476756930352,
+      "num_tokens": 1082903.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6600051839011056,
+      "eval_loss": 0.5857571363449097,
+      "eval_mean_token_accuracy": 0.8248803649629866,
+      "eval_num_tokens": 1101768.0,
+      "eval_runtime": 106.3263,
+      "eval_samples_per_second": 13.12,
+      "eval_steps_per_second": 1.646,
+      "step": 408
+    },
+    {
+      "entropy": 0.5965584144374441,
+      "epoch": 1.1031307550644567,
+      "grad_norm": 0.5178472399711609,
+      "learning_rate": 0.00017488988654901867,
+      "loss": 0.5578182220458985,
+      "mean_token_accuracy": 0.830260409921559,
+      "num_tokens": 1223073.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.5831904768943786,
+      "epoch": 1.2259054634745243,
+      "grad_norm": 0.4383259117603302,
+      "learning_rate": 0.00017467872603358825,
+      "loss": 0.5399176788330078,
+      "mean_token_accuracy": 0.8351450267434121,
+      "num_tokens": 1360978.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.5885515750944614,
+      "epoch": 1.3486801718845918,
+      "grad_norm": 0.5418892502784729,
+      "learning_rate": 0.00017430800739444695,
+      "loss": 0.5402648544311524,
+      "mean_token_accuracy": 0.8344706323742866,
+      "num_tokens": 1492506.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.603163862079382,
+      "epoch": 1.4714548802946594,
+      "grad_norm": 0.5235878825187683,
+      "learning_rate": 0.0001737784089172037,
+      "loss": 0.5534031677246094,
+      "mean_token_accuracy": 0.8304652521014213,
+      "num_tokens": 1622275.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.5745155589282512,
+      "epoch": 1.5942295887047269,
+      "grad_norm": 0.46109989285469055,
+      "learning_rate": 0.00017309089958207074,
+      "loss": 0.5251047134399414,
+      "mean_token_accuracy": 0.8380702176690101,
+      "num_tokens": 1756450.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.5756107035279274,
+      "epoch": 1.7170042971147943,
+      "grad_norm": 0.38209712505340576,
+      "learning_rate": 0.00017224673729096835,
+      "loss": 0.5253059387207031,
+      "mean_token_accuracy": 0.8379077833890914,
+      "num_tokens": 1890634.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.5695902299880982,
+      "epoch": 1.839779005524862,
+      "grad_norm": 0.4063270688056946,
+      "learning_rate": 0.00017124746656600374,
+      "loss": 0.5199185180664062,
+      "mean_token_accuracy": 0.8399301299452782,
+      "num_tokens": 2028046.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.5635769453644752,
+      "epoch": 1.9625537139349294,
+      "grad_norm": 0.4165436029434204,
+      "learning_rate": 0.00017009491572353503,
+      "loss": 0.5156089019775391,
+      "mean_token_accuracy": 0.8404546654224396,
+      "num_tokens": 2163703.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.5587481154714312,
+      "eval_loss": 0.5402795672416687,
+      "eval_mean_token_accuracy": 0.8386626076698304,
+      "eval_num_tokens": 2203536.0,
+      "eval_runtime": 105.9883,
+      "eval_samples_per_second": 13.162,
+      "eval_steps_per_second": 1.651,
+      "step": 816
+    },
+    {
+      "entropy": 0.5002549305785126,
+      "epoch": 2.0834868017188457,
+      "grad_norm": 0.4061429500579834,
+      "learning_rate": 0.00016879119352899107,
+      "loss": 0.4499735641479492,
+      "mean_token_accuracy": 0.8566478448470837,
+      "num_tokens": 2298231.0,
+      "step": 850
+    },
+    {
+      "entropy": 0.4894501394033432,
+      "epoch": 2.2062615101289134,
+      "grad_norm": 0.530322790145874,
+      "learning_rate": 0.00016733868533856737,
+      "loss": 0.4376458740234375,
+      "mean_token_accuracy": 0.859256779551506,
+      "num_tokens": 2427608.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.48188683822751044,
+      "epoch": 2.329036218538981,
+      "grad_norm": 0.4338841438293457,
+      "learning_rate": 0.0001657400487348575,
+      "loss": 0.435361328125,
+      "mean_token_accuracy": 0.8610378202795982,
+      "num_tokens": 2565982.0,
+      "step": 950
+    },
+    {
+      "entropy": 0.4992289599776268,
+      "epoch": 2.4518109269490487,
+      "grad_norm": 0.44097110629081726,
+      "learning_rate": 0.00016399820866440517,
+      "loss": 0.4474017333984375,
+      "mean_token_accuracy": 0.8565557089447975,
+      "num_tokens": 2695545.0,
+      "step": 1000
+    },
+    {
+      "entropy": 0.49334951683878897,
+      "epoch": 2.574585635359116,
+      "grad_norm": 0.48084133863449097,
+      "learning_rate": 0.00016211635208607405,
+      "loss": 0.440751953125,
+      "mean_token_accuracy": 0.8580261027812958,
+      "num_tokens": 2829781.0,
+      "step": 1050
+    },
+    {
+      "entropy": 0.4898635870218277,
+      "epoch": 2.6973603437691835,
+      "grad_norm": 0.40215447545051575,
+      "learning_rate": 0.00016009792214002617,
+      "loss": 0.44588775634765626,
+      "mean_token_accuracy": 0.8572473660111427,
+      "num_tokens": 2970162.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.4932038363814354,
+      "epoch": 2.820135052179251,
+      "grad_norm": 0.4465612769126892,
+      "learning_rate": 0.00015794661184797816,
+      "loss": 0.4444055938720703,
+      "mean_token_accuracy": 0.8583759877085686,
+      "num_tokens": 3103775.0,
+      "step": 1150
+    },
+    {
+      "entropy": 0.4850301502645016,
+      "epoch": 2.942909760589319,
+      "grad_norm": 0.43597331643104553,
+      "learning_rate": 0.0001556663573562617,
+      "loss": 0.4382463073730469,
+      "mean_token_accuracy": 0.8603282216191291,
+      "num_tokens": 3241634.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.4902002043383462,
+      "eval_loss": 0.5276822447776794,
+      "eval_mean_token_accuracy": 0.8454030500139509,
+      "eval_num_tokens": 3305304.0,
+      "eval_runtime": 106.004,
+      "eval_samples_per_second": 13.16,
+      "eval_steps_per_second": 1.651,
+      "step": 1224
+    },
+    {
+      "entropy": 0.4309167733349776,
+      "epoch": 3.063842848373235,
+      "grad_norm": 0.4111804664134979,
+      "learning_rate": 0.00015326133073405068,
+      "loss": 0.3837854766845703,
+      "mean_token_accuracy": 0.8741549949355537,
+      "num_tokens": 3378837.0,
+      "step": 1250
+    },
+    {
+      "entropy": 0.38693042084574697,
+      "epoch": 3.1866175567833026,
+      "grad_norm": 0.46740710735321045,
+      "learning_rate": 0.00015073593233993186,
+      "loss": 0.335601806640625,
+      "mean_token_accuracy": 0.8861287096142769,
+      "num_tokens": 3510318.0,
+      "step": 1300
+    },
+    {
+      "entropy": 0.39316988229751587,
+      "epoch": 3.3093922651933703,
+      "grad_norm": 0.5308303236961365,
+      "learning_rate": 0.00014809478277078602,
+      "loss": 0.34588619232177736,
+      "mean_token_accuracy": 0.8837816697359085,
+      "num_tokens": 3647579.0,
+      "step": 1350
+    },
+    {
+      "entropy": 0.39958367675542833,
+      "epoch": 3.4321669736034375,
+      "grad_norm": 0.5955453515052795,
+      "learning_rate": 0.00014534271440770966,
+      "loss": 0.34634483337402344,
+      "mean_token_accuracy": 0.8832473835349083,
+      "num_tokens": 3779946.0,
+      "step": 1400
+    },
+    {
+      "entropy": 0.4100498601794243,
+      "epoch": 3.554941682013505,
+      "grad_norm": 0.4593118727207184,
+      "learning_rate": 0.0001424847625744456,
+      "loss": 0.3560930633544922,
+      "mean_token_accuracy": 0.8804642978310585,
+      "num_tokens": 3908873.0,
+      "step": 1450
+    },
+    {
+      "entropy": 0.39962531872093676,
+      "epoch": 3.677716390423573,
+      "grad_norm": 0.4584615230560303,
+      "learning_rate": 0.00013952615632449963,
+      "loss": 0.34978450775146486,
+      "mean_token_accuracy": 0.8825609818100929,
+      "num_tokens": 4046620.0,
+      "step": 1500
+    },
+    {
+      "entropy": 0.39326138779520986,
+      "epoch": 3.80049109883364,
+      "grad_norm": 0.5227728486061096,
+      "learning_rate": 0.00013647230887379902,
+      "loss": 0.3470985794067383,
+      "mean_token_accuracy": 0.8833131140470505,
+      "num_tokens": 4184429.0,
+      "step": 1550
+    },
+    {
+      "entropy": 0.399793985709548,
+      "epoch": 3.9232658072437077,
+      "grad_norm": 0.4263511002063751,
+      "learning_rate": 0.00013332880769639856,
+      "loss": 0.3550622177124023,
+      "mean_token_accuracy": 0.8825028780102729,
+      "num_tokens": 4318596.0,
+      "step": 1600
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.41604454040527344,
+      "eval_loss": 0.5501811504364014,
+      "eval_mean_token_accuracy": 0.8466314860752651,
+      "eval_num_tokens": 4407072.0,
+      "eval_runtime": 105.9531,
+      "eval_samples_per_second": 13.166,
+      "eval_steps_per_second": 1.652,
+      "step": 1632
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4080,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.365209251156378e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-2040/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-2040/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 256,
+  "lora_bias": false,
+  "lora_dropout": 0.08935122001117755,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 128,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "o_proj",
+    "k_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj",
+    "down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-2040/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-2040/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-2040/trainer_state.json

@@ -0,0 +1,489 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 2040,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.4665306943655014,
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.8344842791557312,
+      "learning_rate": 2.1010394822962346e-05,
+      "loss": 1.379196014404297,
+      "mean_token_accuracy": 0.6848452094197274,
+      "num_tokens": 137019.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.7834547007083893,
+      "epoch": 0.24554941682013506,
+      "grad_norm": 1.1248278617858887,
+      "learning_rate": 4.2449573213740254e-05,
+      "loss": 0.7345225524902343,
+      "mean_token_accuracy": 0.7882961884140969,
+      "num_tokens": 267448.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.7194462446868419,
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.6498261094093323,
+      "learning_rate": 6.388875160451815e-05,
+      "loss": 0.664264907836914,
+      "mean_token_accuracy": 0.8042438223958015,
+      "num_tokens": 409071.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.6771699503064156,
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.6653317809104919,
+      "learning_rate": 8.532792999529606e-05,
+      "loss": 0.6322076797485352,
+      "mean_token_accuracy": 0.8130025643110276,
+      "num_tokens": 542207.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.6632554066181183,
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.44293469190597534,
+      "learning_rate": 0.00010676710838607398,
+      "loss": 0.617110595703125,
+      "mean_token_accuracy": 0.8168379080295562,
+      "num_tokens": 679619.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.6508415986597538,
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.49295490980148315,
+      "learning_rate": 0.00012820628677685187,
+      "loss": 0.6023779296875,
+      "mean_token_accuracy": 0.8187075120210647,
+      "num_tokens": 815397.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.6380941693484783,
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.5855488181114197,
+      "learning_rate": 0.00014964546516762976,
+      "loss": 0.5912636947631836,
+      "mean_token_accuracy": 0.8227647486329078,
+      "num_tokens": 949315.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.6346297286450863,
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.3817698657512665,
+      "learning_rate": 0.00017108464355840766,
+      "loss": 0.5902091598510742,
+      "mean_token_accuracy": 0.8231476756930352,
+      "num_tokens": 1082903.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6600051839011056,
+      "eval_loss": 0.5857571363449097,
+      "eval_mean_token_accuracy": 0.8248803649629866,
+      "eval_num_tokens": 1101768.0,
+      "eval_runtime": 106.3263,
+      "eval_samples_per_second": 13.12,
+      "eval_steps_per_second": 1.646,
+      "step": 408
+    },
+    {
+      "entropy": 0.5965584144374441,
+      "epoch": 1.1031307550644567,
+      "grad_norm": 0.5178472399711609,
+      "learning_rate": 0.00017488988654901867,
+      "loss": 0.5578182220458985,
+      "mean_token_accuracy": 0.830260409921559,
+      "num_tokens": 1223073.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.5831904768943786,
+      "epoch": 1.2259054634745243,
+      "grad_norm": 0.4383259117603302,
+      "learning_rate": 0.00017467872603358825,
+      "loss": 0.5399176788330078,
+      "mean_token_accuracy": 0.8351450267434121,
+      "num_tokens": 1360978.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.5885515750944614,
+      "epoch": 1.3486801718845918,
+      "grad_norm": 0.5418892502784729,
+      "learning_rate": 0.00017430800739444695,
+      "loss": 0.5402648544311524,
+      "mean_token_accuracy": 0.8344706323742866,
+      "num_tokens": 1492506.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.603163862079382,
+      "epoch": 1.4714548802946594,
+      "grad_norm": 0.5235878825187683,
+      "learning_rate": 0.0001737784089172037,
+      "loss": 0.5534031677246094,
+      "mean_token_accuracy": 0.8304652521014213,
+      "num_tokens": 1622275.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.5745155589282512,
+      "epoch": 1.5942295887047269,
+      "grad_norm": 0.46109989285469055,
+      "learning_rate": 0.00017309089958207074,
+      "loss": 0.5251047134399414,
+      "mean_token_accuracy": 0.8380702176690101,
+      "num_tokens": 1756450.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.5756107035279274,
+      "epoch": 1.7170042971147943,
+      "grad_norm": 0.38209712505340576,
+      "learning_rate": 0.00017224673729096835,
+      "loss": 0.5253059387207031,
+      "mean_token_accuracy": 0.8379077833890914,
+      "num_tokens": 1890634.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.5695902299880982,
+      "epoch": 1.839779005524862,
+      "grad_norm": 0.4063270688056946,
+      "learning_rate": 0.00017124746656600374,
+      "loss": 0.5199185180664062,
+      "mean_token_accuracy": 0.8399301299452782,
+      "num_tokens": 2028046.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.5635769453644752,
+      "epoch": 1.9625537139349294,
+      "grad_norm": 0.4165436029434204,
+      "learning_rate": 0.00017009491572353503,
+      "loss": 0.5156089019775391,
+      "mean_token_accuracy": 0.8404546654224396,
+      "num_tokens": 2163703.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.5587481154714312,
+      "eval_loss": 0.5402795672416687,
+      "eval_mean_token_accuracy": 0.8386626076698304,
+      "eval_num_tokens": 2203536.0,
+      "eval_runtime": 105.9883,
+      "eval_samples_per_second": 13.162,
+      "eval_steps_per_second": 1.651,
+      "step": 816
+    },
+    {
+      "entropy": 0.5002549305785126,
+      "epoch": 2.0834868017188457,
+      "grad_norm": 0.4061429500579834,
+      "learning_rate": 0.00016879119352899107,
+      "loss": 0.4499735641479492,
+      "mean_token_accuracy": 0.8566478448470837,
+      "num_tokens": 2298231.0,
+      "step": 850
+    },
+    {
+      "entropy": 0.4894501394033432,
+      "epoch": 2.2062615101289134,
+      "grad_norm": 0.530322790145874,
+      "learning_rate": 0.00016733868533856737,
+      "loss": 0.4376458740234375,
+      "mean_token_accuracy": 0.859256779551506,
+      "num_tokens": 2427608.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.48188683822751044,
+      "epoch": 2.329036218538981,
+      "grad_norm": 0.4338841438293457,
+      "learning_rate": 0.0001657400487348575,
+      "loss": 0.435361328125,
+      "mean_token_accuracy": 0.8610378202795982,
+      "num_tokens": 2565982.0,
+      "step": 950
+    },
+    {
+      "entropy": 0.4992289599776268,
+      "epoch": 2.4518109269490487,
+      "grad_norm": 0.44097110629081726,
+      "learning_rate": 0.00016399820866440517,
+      "loss": 0.4474017333984375,
+      "mean_token_accuracy": 0.8565557089447975,
+      "num_tokens": 2695545.0,
+      "step": 1000
+    },
+    {
+      "entropy": 0.49334951683878897,
+      "epoch": 2.574585635359116,
+      "grad_norm": 0.48084133863449097,
+      "learning_rate": 0.00016211635208607405,
+      "loss": 0.440751953125,
+      "mean_token_accuracy": 0.8580261027812958,
+      "num_tokens": 2829781.0,
+      "step": 1050
+    },
+    {
+      "entropy": 0.4898635870218277,
+      "epoch": 2.6973603437691835,
+      "grad_norm": 0.40215447545051575,
+      "learning_rate": 0.00016009792214002617,
+      "loss": 0.44588775634765626,
+      "mean_token_accuracy": 0.8572473660111427,
+      "num_tokens": 2970162.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.4932038363814354,
+      "epoch": 2.820135052179251,
+      "grad_norm": 0.4465612769126892,
+      "learning_rate": 0.00015794661184797816,
+      "loss": 0.4444055938720703,
+      "mean_token_accuracy": 0.8583759877085686,
+      "num_tokens": 3103775.0,
+      "step": 1150
+    },
+    {
+      "entropy": 0.4850301502645016,
+      "epoch": 2.942909760589319,
+      "grad_norm": 0.43597331643104553,
+      "learning_rate": 0.0001556663573562617,
+      "loss": 0.4382463073730469,
+      "mean_token_accuracy": 0.8603282216191291,
+      "num_tokens": 3241634.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.4902002043383462,
+      "eval_loss": 0.5276822447776794,
+      "eval_mean_token_accuracy": 0.8454030500139509,
+      "eval_num_tokens": 3305304.0,
+      "eval_runtime": 106.004,
+      "eval_samples_per_second": 13.16,
+      "eval_steps_per_second": 1.651,
+      "step": 1224
+    },
+    {
+      "entropy": 0.4309167733349776,
+      "epoch": 3.063842848373235,
+      "grad_norm": 0.4111804664134979,
+      "learning_rate": 0.00015326133073405068,
+      "loss": 0.3837854766845703,
+      "mean_token_accuracy": 0.8741549949355537,
+      "num_tokens": 3378837.0,
+      "step": 1250
+    },
+    {
+      "entropy": 0.38693042084574697,
+      "epoch": 3.1866175567833026,
+      "grad_norm": 0.46740710735321045,
+      "learning_rate": 0.00015073593233993186,
+      "loss": 0.335601806640625,
+      "mean_token_accuracy": 0.8861287096142769,
+      "num_tokens": 3510318.0,
+      "step": 1300
+    },
+    {
+      "entropy": 0.39316988229751587,
+      "epoch": 3.3093922651933703,
+      "grad_norm": 0.5308303236961365,
+      "learning_rate": 0.00014809478277078602,
+      "loss": 0.34588619232177736,
+      "mean_token_accuracy": 0.8837816697359085,
+      "num_tokens": 3647579.0,
+      "step": 1350
+    },
+    {
+      "entropy": 0.39958367675542833,
+      "epoch": 3.4321669736034375,
+      "grad_norm": 0.5955453515052795,
+      "learning_rate": 0.00014534271440770966,
+      "loss": 0.34634483337402344,
+      "mean_token_accuracy": 0.8832473835349083,
+      "num_tokens": 3779946.0,
+      "step": 1400
+    },
+    {
+      "entropy": 0.4100498601794243,
+      "epoch": 3.554941682013505,
+      "grad_norm": 0.4593118727207184,
+      "learning_rate": 0.0001424847625744456,
+      "loss": 0.3560930633544922,
+      "mean_token_accuracy": 0.8804642978310585,
+      "num_tokens": 3908873.0,
+      "step": 1450
+    },
+    {
+      "entropy": 0.39962531872093676,
+      "epoch": 3.677716390423573,
+      "grad_norm": 0.4584615230560303,
+      "learning_rate": 0.00013952615632449963,
+      "loss": 0.34978450775146486,
+      "mean_token_accuracy": 0.8825609818100929,
+      "num_tokens": 4046620.0,
+      "step": 1500
+    },
+    {
+      "entropy": 0.39326138779520986,
+      "epoch": 3.80049109883364,
+      "grad_norm": 0.5227728486061096,
+      "learning_rate": 0.00013647230887379902,
+      "loss": 0.3470985794067383,
+      "mean_token_accuracy": 0.8833131140470505,
+      "num_tokens": 4184429.0,
+      "step": 1550
+    },
+    {
+      "entropy": 0.399793985709548,
+      "epoch": 3.9232658072437077,
+      "grad_norm": 0.4263511002063751,
+      "learning_rate": 0.00013332880769639856,
+      "loss": 0.3550622177124023,
+      "mean_token_accuracy": 0.8825028780102729,
+      "num_tokens": 4318596.0,
+      "step": 1600
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.41604454040527344,
+      "eval_loss": 0.5501811504364014,
+      "eval_mean_token_accuracy": 0.8466314860752651,
+      "eval_num_tokens": 4407072.0,
+      "eval_runtime": 105.9531,
+      "eval_samples_per_second": 13.166,
+      "eval_steps_per_second": 1.652,
+      "step": 1632
+    },
+    {
+      "entropy": 0.3492097145107192,
+      "epoch": 4.044198895027624,
+      "grad_norm": 0.41719651222229004,
+      "learning_rate": 0.00013010140430135466,
+      "loss": 0.30388387680053713,
+      "mean_token_accuracy": 0.8972433387930623,
+      "num_tokens": 4455549.0,
+      "step": 1650
+    },
+    {
+      "entropy": 0.2980811195075512,
+      "epoch": 4.166973603437691,
+      "grad_norm": 0.694450318813324,
+      "learning_rate": 0.00012679600370947314,
+      "loss": 0.24677230834960937,
+      "mean_token_accuracy": 0.9134943392872811,
+      "num_tokens": 4584218.0,
+      "step": 1700
+    },
+    {
+      "entropy": 0.2911450309306383,
+      "epoch": 4.2897483118477595,
+      "grad_norm": 0.4893203675746918,
+      "learning_rate": 0.00012341865364918372,
+      "loss": 0.24411561965942383,
+      "mean_token_accuracy": 0.9145414170622825,
+      "num_tokens": 4724425.0,
+      "step": 1750
+    },
+    {
+      "entropy": 0.2982170185446739,
+      "epoch": 4.412523020257827,
+      "grad_norm": 0.5446121096611023,
+      "learning_rate": 0.00011997553349130966,
+      "loss": 0.24687444686889648,
+      "mean_token_accuracy": 0.9126561579108238,
+      "num_tokens": 4861552.0,
+      "step": 1800
+    },
+    {
+      "entropy": 0.30126186162233354,
+      "epoch": 4.535297728667894,
+      "grad_norm": 0.5465192198753357,
+      "learning_rate": 0.00011647294294297787,
+      "loss": 0.2527424621582031,
+      "mean_token_accuracy": 0.9110122618079185,
+      "num_tokens": 4993610.0,
+      "step": 1850
+    },
+    {
+      "entropy": 0.2993039993941784,
+      "epoch": 4.658072437077962,
+      "grad_norm": 0.6376634836196899,
+      "learning_rate": 0.00011291729052135527,
+      "loss": 0.2507955169677734,
+      "mean_token_accuracy": 0.9124656105041504,
+      "num_tokens": 5126291.0,
+      "step": 1900
+    },
+    {
+      "entropy": 0.29630616143345834,
+      "epoch": 4.780847145488029,
+      "grad_norm": 0.5191586017608643,
+      "learning_rate": 0.00010931508182830125,
+      "loss": 0.2503428077697754,
+      "mean_token_accuracy": 0.9127612486481667,
+      "num_tokens": 5266076.0,
+      "step": 1950
+    },
+    {
+      "entropy": 0.29707231901586056,
+      "epoch": 4.903621853898097,
+      "grad_norm": 0.42329415678977966,
+      "learning_rate": 0.00010567290764738874,
+      "loss": 0.24930551528930664,
+      "mean_token_accuracy": 0.911863605082035,
+      "num_tokens": 5403761.0,
+      "step": 2000
+    },
+    {
+      "epoch": 5.0,
+      "eval_entropy": 0.34850709250995093,
+      "eval_loss": 0.6053183078765869,
+      "eval_mean_token_accuracy": 0.8461176470347813,
+      "eval_num_tokens": 5508840.0,
+      "eval_runtime": 105.9653,
+      "eval_samples_per_second": 13.165,
+      "eval_steps_per_second": 1.651,
+      "step": 2040
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4080,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.203960855839949e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-2448/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-2448/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 256,
+  "lora_bias": false,
+  "lora_dropout": 0.08935122001117755,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 128,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "o_proj",
+    "k_proj",
+    "q_proj",
+    "up_proj",
+    "v_proj",
+    "down_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-2448/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

random_original_Estonian/Qwen3-14B-Base_random_splits_original_features_train_random_splits_original_features_test2/checkpoint-2448/trainer_state.json

@@ -0,0 +1,580 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 6.0,
+  "eval_steps": 500,
+  "global_step": 2448,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.4665306943655014,
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.8344842791557312,
+      "learning_rate": 2.1010394822962346e-05,
+      "loss": 1.379196014404297,
+      "mean_token_accuracy": 0.6848452094197274,
+      "num_tokens": 137019.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.7834547007083893,
+      "epoch": 0.24554941682013506,
+      "grad_norm": 1.1248278617858887,
+      "learning_rate": 4.2449573213740254e-05,
+      "loss": 0.7345225524902343,
+      "mean_token_accuracy": 0.7882961884140969,
+      "num_tokens": 267448.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.7194462446868419,
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.6498261094093323,
+      "learning_rate": 6.388875160451815e-05,
+      "loss": 0.664264907836914,
+      "mean_token_accuracy": 0.8042438223958015,
+      "num_tokens": 409071.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.6771699503064156,
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.6653317809104919,
+      "learning_rate": 8.532792999529606e-05,
+      "loss": 0.6322076797485352,
+      "mean_token_accuracy": 0.8130025643110276,
+      "num_tokens": 542207.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.6632554066181183,
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.44293469190597534,
+      "learning_rate": 0.00010676710838607398,
+      "loss": 0.617110595703125,
+      "mean_token_accuracy": 0.8168379080295562,
+      "num_tokens": 679619.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.6508415986597538,
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.49295490980148315,
+      "learning_rate": 0.00012820628677685187,
+      "loss": 0.6023779296875,
+      "mean_token_accuracy": 0.8187075120210647,
+      "num_tokens": 815397.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.6380941693484783,
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.5855488181114197,
+      "learning_rate": 0.00014964546516762976,
+      "loss": 0.5912636947631836,
+      "mean_token_accuracy": 0.8227647486329078,
+      "num_tokens": 949315.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.6346297286450863,
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.3817698657512665,
+      "learning_rate": 0.00017108464355840766,
+      "loss": 0.5902091598510742,
+      "mean_token_accuracy": 0.8231476756930352,
+      "num_tokens": 1082903.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6600051839011056,
+      "eval_loss": 0.5857571363449097,
+      "eval_mean_token_accuracy": 0.8248803649629866,
+      "eval_num_tokens": 1101768.0,
+      "eval_runtime": 106.3263,
+      "eval_samples_per_second": 13.12,
+      "eval_steps_per_second": 1.646,
+      "step": 408
+    },
+    {
+      "entropy": 0.5965584144374441,
+      "epoch": 1.1031307550644567,
+      "grad_norm": 0.5178472399711609,
+      "learning_rate": 0.00017488988654901867,
+      "loss": 0.5578182220458985,
+      "mean_token_accuracy": 0.830260409921559,
+      "num_tokens": 1223073.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.5831904768943786,
+      "epoch": 1.2259054634745243,
+      "grad_norm": 0.4383259117603302,
+      "learning_rate": 0.00017467872603358825,
+      "loss": 0.5399176788330078,
+      "mean_token_accuracy": 0.8351450267434121,
+      "num_tokens": 1360978.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.5885515750944614,
+      "epoch": 1.3486801718845918,
+      "grad_norm": 0.5418892502784729,
+      "learning_rate": 0.00017430800739444695,
+      "loss": 0.5402648544311524,
+      "mean_token_accuracy": 0.8344706323742866,
+      "num_tokens": 1492506.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.603163862079382,
+      "epoch": 1.4714548802946594,
+      "grad_norm": 0.5235878825187683,
+      "learning_rate": 0.0001737784089172037,
+      "loss": 0.5534031677246094,
+      "mean_token_accuracy": 0.8304652521014213,
+      "num_tokens": 1622275.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.5745155589282512,
+      "epoch": 1.5942295887047269,
+      "grad_norm": 0.46109989285469055,
+      "learning_rate": 0.00017309089958207074,
+      "loss": 0.5251047134399414,
+      "mean_token_accuracy": 0.8380702176690101,
+      "num_tokens": 1756450.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.5756107035279274,
+      "epoch": 1.7170042971147943,
+      "grad_norm": 0.38209712505340576,
+      "learning_rate": 0.00017224673729096835,
+      "loss": 0.5253059387207031,
+      "mean_token_accuracy": 0.8379077833890914,
+      "num_tokens": 1890634.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.5695902299880982,
+      "epoch": 1.839779005524862,
+      "grad_norm": 0.4063270688056946,
+      "learning_rate": 0.00017124746656600374,
+      "loss": 0.5199185180664062,
+      "mean_token_accuracy": 0.8399301299452782,
+      "num_tokens": 2028046.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.5635769453644752,
+      "epoch": 1.9625537139349294,
+      "grad_norm": 0.4165436029434204,
+      "learning_rate": 0.00017009491572353503,
+      "loss": 0.5156089019775391,
+      "mean_token_accuracy": 0.8404546654224396,
+      "num_tokens": 2163703.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.5587481154714312,
+      "eval_loss": 0.5402795672416687,
+      "eval_mean_token_accuracy": 0.8386626076698304,
+      "eval_num_tokens": 2203536.0,
+      "eval_runtime": 105.9883,
+      "eval_samples_per_second": 13.162,
+      "eval_steps_per_second": 1.651,
+      "step": 816
+    },
+    {
+      "entropy": 0.5002549305785126,
+      "epoch": 2.0834868017188457,
+      "grad_norm": 0.4061429500579834,
+      "learning_rate": 0.00016879119352899107,
+      "loss": 0.4499735641479492,
+      "mean_token_accuracy": 0.8566478448470837,
+      "num_tokens": 2298231.0,
+      "step": 850
+    },
+    {
+      "entropy": 0.4894501394033432,
+      "epoch": 2.2062615101289134,
+      "grad_norm": 0.530322790145874,
+      "learning_rate": 0.00016733868533856737,
+      "loss": 0.4376458740234375,
+      "mean_token_accuracy": 0.859256779551506,
+      "num_tokens": 2427608.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.48188683822751044,
+      "epoch": 2.329036218538981,
+      "grad_norm": 0.4338841438293457,
+      "learning_rate": 0.0001657400487348575,
+      "loss": 0.435361328125,
+      "mean_token_accuracy": 0.8610378202795982,
+      "num_tokens": 2565982.0,
+      "step": 950
+    },
+    {
+      "entropy": 0.4992289599776268,
+      "epoch": 2.4518109269490487,
+      "grad_norm": 0.44097110629081726,
+      "learning_rate": 0.00016399820866440517,
+      "loss": 0.4474017333984375,
+      "mean_token_accuracy": 0.8565557089447975,
+      "num_tokens": 2695545.0,
+      "step": 1000
+    },
+    {
+      "entropy": 0.49334951683878897,
+      "epoch": 2.574585635359116,
+      "grad_norm": 0.48084133863449097,
+      "learning_rate": 0.00016211635208607405,
+      "loss": 0.440751953125,
+      "mean_token_accuracy": 0.8580261027812958,
+      "num_tokens": 2829781.0,
+      "step": 1050
+    },
+    {
+      "entropy": 0.4898635870218277,
+      "epoch": 2.6973603437691835,
+      "grad_norm": 0.40215447545051575,
+      "learning_rate": 0.00016009792214002617,
+      "loss": 0.44588775634765626,
+      "mean_token_accuracy": 0.8572473660111427,
+      "num_tokens": 2970162.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.4932038363814354,
+      "epoch": 2.820135052179251,
+      "grad_norm": 0.4465612769126892,
+      "learning_rate": 0.00015794661184797816,
+      "loss": 0.4444055938720703,
+      "mean_token_accuracy": 0.8583759877085686,
+      "num_tokens": 3103775.0,
+      "step": 1150
+    },
+    {
+      "entropy": 0.4850301502645016,
+      "epoch": 2.942909760589319,
+      "grad_norm": 0.43597331643104553,
+      "learning_rate": 0.0001556663573562617,
+      "loss": 0.4382463073730469,
+      "mean_token_accuracy": 0.8603282216191291,
+      "num_tokens": 3241634.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.4902002043383462,
+      "eval_loss": 0.5276822447776794,
+      "eval_mean_token_accuracy": 0.8454030500139509,
+      "eval_num_tokens": 3305304.0,
+      "eval_runtime": 106.004,
+      "eval_samples_per_second": 13.16,
+      "eval_steps_per_second": 1.651,
+      "step": 1224
+    },
+    {
+      "entropy": 0.4309167733349776,
+      "epoch": 3.063842848373235,
+      "grad_norm": 0.4111804664134979,
+      "learning_rate": 0.00015326133073405068,
+      "loss": 0.3837854766845703,
+      "mean_token_accuracy": 0.8741549949355537,
+      "num_tokens": 3378837.0,
+      "step": 1250
+    },
+    {
+      "entropy": 0.38693042084574697,
+      "epoch": 3.1866175567833026,
+      "grad_norm": 0.46740710735321045,
+      "learning_rate": 0.00015073593233993186,
+      "loss": 0.335601806640625,
+      "mean_token_accuracy": 0.8861287096142769,
+      "num_tokens": 3510318.0,
+      "step": 1300
+    },
+    {
+      "entropy": 0.39316988229751587,
+      "epoch": 3.3093922651933703,
+      "grad_norm": 0.5308303236961365,
+      "learning_rate": 0.00014809478277078602,
+      "loss": 0.34588619232177736,
+      "mean_token_accuracy": 0.8837816697359085,
+      "num_tokens": 3647579.0,
+      "step": 1350
+    },
+    {
+      "entropy": 0.39958367675542833,
+      "epoch": 3.4321669736034375,
+      "grad_norm": 0.5955453515052795,
+      "learning_rate": 0.00014534271440770966,
+      "loss": 0.34634483337402344,
+      "mean_token_accuracy": 0.8832473835349083,
+      "num_tokens": 3779946.0,
+      "step": 1400
+    },
+    {
+      "entropy": 0.4100498601794243,
+      "epoch": 3.554941682013505,
+      "grad_norm": 0.4593118727207184,
+      "learning_rate": 0.0001424847625744456,
+      "loss": 0.3560930633544922,
+      "mean_token_accuracy": 0.8804642978310585,
+      "num_tokens": 3908873.0,
+      "step": 1450
+    },
+    {
+      "entropy": 0.39962531872093676,
+      "epoch": 3.677716390423573,
+      "grad_norm": 0.4584615230560303,
+      "learning_rate": 0.00013952615632449963,
+      "loss": 0.34978450775146486,
+      "mean_token_accuracy": 0.8825609818100929,
+      "num_tokens": 4046620.0,
+      "step": 1500
+    },
+    {
+      "entropy": 0.39326138779520986,
+      "epoch": 3.80049109883364,
+      "grad_norm": 0.5227728486061096,
+      "learning_rate": 0.00013647230887379902,
+      "loss": 0.3470985794067383,
+      "mean_token_accuracy": 0.8833131140470505,
+      "num_tokens": 4184429.0,
+      "step": 1550
+    },
+    {
+      "entropy": 0.399793985709548,
+      "epoch": 3.9232658072437077,
+      "grad_norm": 0.4263511002063751,
+      "learning_rate": 0.00013332880769639856,
+      "loss": 0.3550622177124023,
+      "mean_token_accuracy": 0.8825028780102729,
+      "num_tokens": 4318596.0,
+      "step": 1600
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.41604454040527344,
+      "eval_loss": 0.5501811504364014,
+      "eval_mean_token_accuracy": 0.8466314860752651,
+      "eval_num_tokens": 4407072.0,
+      "eval_runtime": 105.9531,
+      "eval_samples_per_second": 13.166,
+      "eval_steps_per_second": 1.652,
+      "step": 1632
+    },
+    {
+      "entropy": 0.3492097145107192,
+      "epoch": 4.044198895027624,
+      "grad_norm": 0.41719651222229004,
+      "learning_rate": 0.00013010140430135466,
+      "loss": 0.30388387680053713,
+      "mean_token_accuracy": 0.8972433387930623,
+      "num_tokens": 4455549.0,
+      "step": 1650
+    },
+    {
+      "entropy": 0.2980811195075512,
+      "epoch": 4.166973603437691,
+      "grad_norm": 0.694450318813324,
+      "learning_rate": 0.00012679600370947314,
+      "loss": 0.24677230834960937,
+      "mean_token_accuracy": 0.9134943392872811,
+      "num_tokens": 4584218.0,
+      "step": 1700
+    },
+    {
+      "entropy": 0.2911450309306383,
+      "epoch": 4.2897483118477595,
+      "grad_norm": 0.4893203675746918,
+      "learning_rate": 0.00012341865364918372,
+      "loss": 0.24411561965942383,
+      "mean_token_accuracy": 0.9145414170622825,
+      "num_tokens": 4724425.0,
+      "step": 1750
+    },
+    {
+      "entropy": 0.2982170185446739,
+      "epoch": 4.412523020257827,
+      "grad_norm": 0.5446121096611023,
+      "learning_rate": 0.00011997553349130966,
+      "loss": 0.24687444686889648,
+      "mean_token_accuracy": 0.9126561579108238,
+      "num_tokens": 4861552.0,
+      "step": 1800
+    },
+    {
+      "entropy": 0.30126186162233354,
+      "epoch": 4.535297728667894,
+      "grad_norm": 0.5465192198753357,
+      "learning_rate": 0.00011647294294297787,
+      "loss": 0.2527424621582031,
+      "mean_token_accuracy": 0.9110122618079185,
+      "num_tokens": 4993610.0,
+      "step": 1850
+    },
+    {
+      "entropy": 0.2993039993941784,
+      "epoch": 4.658072437077962,
+      "grad_norm": 0.6376634836196899,
+      "learning_rate": 0.00011291729052135527,
+      "loss": 0.2507955169677734,
+      "mean_token_accuracy": 0.9124656105041504,
+      "num_tokens": 5126291.0,
+      "step": 1900
+    },
+    {
+      "entropy": 0.29630616143345834,
+      "epoch": 4.780847145488029,
+      "grad_norm": 0.5191586017608643,
+      "learning_rate": 0.00010931508182830125,
+      "loss": 0.2503428077697754,
+      "mean_token_accuracy": 0.9127612486481667,
+      "num_tokens": 5266076.0,
+      "step": 1950
+    },
+    {
+      "entropy": 0.29707231901586056,
+      "epoch": 4.903621853898097,
+      "grad_norm": 0.42329415678977966,
+      "learning_rate": 0.00010567290764738874,
+      "loss": 0.24930551528930664,
+      "mean_token_accuracy": 0.911863605082035,
+      "num_tokens": 5403761.0,
+      "step": 2000
+    },
+    {
+      "epoch": 5.0,
+      "eval_entropy": 0.34850709250995093,
+      "eval_loss": 0.6053183078765869,
+      "eval_mean_token_accuracy": 0.8461176470347813,
+      "eval_num_tokens": 5508840.0,
+      "eval_runtime": 105.9653,
+      "eval_samples_per_second": 13.165,
+      "eval_steps_per_second": 1.651,
+      "step": 2040
+    },
+    {
+      "entropy": 0.28518860919523964,
+      "epoch": 5.024554941682013,
+      "grad_norm": 0.6316563487052917,
+      "learning_rate": 0.00010199743188507247,
+      "loss": 0.23341741561889648,
+      "mean_token_accuracy": 0.9183166952907736,
+      "num_tokens": 5535743.0,
+      "step": 2050
+    },
+    {
+      "entropy": 0.20185263987630606,
+      "epoch": 5.147329650092081,
+      "grad_norm": 0.6542733311653137,
+      "learning_rate": 9.829537937806829e-05,
+      "loss": 0.1496055793762207,
+      "mean_token_accuracy": 0.9459337556362152,
+      "num_tokens": 5675294.0,
+      "step": 2100
+    },
+    {
+      "entropy": 0.2101737355440855,
+      "epoch": 5.270104358502149,
+      "grad_norm": 0.46600785851478577,
+      "learning_rate": 9.457352358925075e-05,
+      "loss": 0.15642745018005372,
+      "mean_token_accuracy": 0.944338481426239,
+      "num_tokens": 5808279.0,
+      "step": 2150
+    },
+    {
+      "entropy": 0.200700718536973,
+      "epoch": 5.392879066912216,
+      "grad_norm": 0.5562126636505127,
+      "learning_rate": 9.083867421458257e-05,
+      "loss": 0.15122636795043945,
+      "mean_token_accuracy": 0.9447117391228675,
+      "num_tokens": 5948774.0,
+      "step": 2200
+    },
+    {
+      "entropy": 0.2139152016490698,
+      "epoch": 5.515653775322283,
+      "grad_norm": 0.5136730670928955,
+      "learning_rate": 8.709766472374944e-05,
+      "loss": 0.16092844009399415,
+      "mean_token_accuracy": 0.9423242849111557,
+      "num_tokens": 6081211.0,
+      "step": 2250
+    },
+    {
+      "entropy": 0.20662450045347214,
+      "epoch": 5.638428483732351,
+      "grad_norm": 0.6007277965545654,
+      "learning_rate": 8.335733985729791e-05,
+      "loss": 0.15495609283447265,
+      "mean_token_accuracy": 0.9431806969642639,
+      "num_tokens": 6213793.0,
+      "step": 2300
+    },
+    {
+      "entropy": 0.20562831707298757,
+      "epoch": 5.7612031921424185,
+      "grad_norm": 0.5216547846794128,
+      "learning_rate": 7.962454310315095e-05,
+      "loss": 0.1568380641937256,
+      "mean_token_accuracy": 0.9425676354765892,
+      "num_tokens": 6350388.0,
+      "step": 2350
+    },
+    {
+      "entropy": 0.20440548222512006,
+      "epoch": 5.883977900552486,
+      "grad_norm": 0.5724782943725586,
+      "learning_rate": 7.590610417541581e-05,
+      "loss": 0.15627790451049806,
+      "mean_token_accuracy": 0.9437827947735786,
+      "num_tokens": 6485960.0,
+      "step": 2400
+    },
+    {
+      "epoch": 6.0,
+      "eval_entropy": 0.2856926244497299,
+      "eval_loss": 0.7071177959442139,
+      "eval_mean_token_accuracy": 0.8432521397726876,
+      "eval_num_tokens": 6610608.0,
+      "eval_runtime": 105.9542,
+      "eval_samples_per_second": 13.166,
+      "eval_steps_per_second": 1.652,
+      "step": 2448
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4080,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.1042614122067046e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 0.005046715367502997,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "up_proj",
+    "k_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1224/trainer_state.json

@@ -0,0 +1,340 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1224,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.8492428332567215,
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.5106982588768005,
+      "learning_rate": 3.492278846049215e-05,
+      "loss": 1.7872950744628906,
+      "mean_token_accuracy": 0.6242395298182964,
+      "num_tokens": 141525.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.8940915194153786,
+      "epoch": 0.24554941682013506,
+      "grad_norm": 0.6675613522529602,
+      "learning_rate": 7.055828688956577e-05,
+      "loss": 0.8397601318359375,
+      "mean_token_accuracy": 0.7503150171041488,
+      "num_tokens": 276389.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.7644237650930882,
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.43234843015670776,
+      "learning_rate": 0.00010619378531863939,
+      "loss": 0.7161208343505859,
+      "mean_token_accuracy": 0.778390144109726,
+      "num_tokens": 422723.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.7164452224969864,
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.4049525558948517,
+      "learning_rate": 0.00014182928374771302,
+      "loss": 0.6739736938476563,
+      "mean_token_accuracy": 0.7887251850962639,
+      "num_tokens": 560353.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.7021449965238571,
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.3455510139465332,
+      "learning_rate": 0.00017746478217678664,
+      "loss": 0.6584578704833984,
+      "mean_token_accuracy": 0.7932635286450386,
+      "num_tokens": 702231.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.6882825265824795,
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.30983075499534607,
+      "learning_rate": 0.00021310028060586025,
+      "loss": 0.6412281799316406,
+      "mean_token_accuracy": 0.7957331365346909,
+      "num_tokens": 842451.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.6746629500389099,
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.4235651195049286,
+      "learning_rate": 0.0002487357790349339,
+      "loss": 0.6273543167114258,
+      "mean_token_accuracy": 0.8004380378127098,
+      "num_tokens": 980824.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.6693046718835831,
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.2910408079624176,
+      "learning_rate": 0.0002843712774640075,
+      "loss": 0.6247041320800781,
+      "mean_token_accuracy": 0.8027204623818398,
+      "num_tokens": 1118697.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6973312156541007,
+      "eval_mean_token_accuracy": 0.7841839187485831,
+      "eval_not_syn_loss": 0.6392670273780823,
+      "eval_not_syn_runtime": 106.0995,
+      "eval_not_syn_samples_per_second": 13.148,
+      "eval_not_syn_steps_per_second": 1.649,
+      "eval_num_tokens": 1138193.0,
+      "step": 408
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6523867055347987,
+      "eval_mean_token_accuracy": 0.8224203096117292,
+      "eval_num_tokens": 1138193.0,
+      "eval_syn_loss": 0.5988379120826721,
+      "eval_syn_runtime": 112.4432,
+      "eval_syn_samples_per_second": 12.406,
+      "eval_syn_steps_per_second": 1.556,
+      "step": 408
+    },
+    {
+      "entropy": 0.6396199787026129,
+      "epoch": 1.1031307550644567,
+      "grad_norm": 0.3523969054222107,
+      "learning_rate": 0.0002906962274290906,
+      "loss": 0.5951201248168946,
+      "mean_token_accuracy": 0.809070566584011,
+      "num_tokens": 1263520.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.6200365445017815,
+      "epoch": 1.2259054634745243,
+      "grad_norm": 0.28092390298843384,
+      "learning_rate": 0.00029034524335317383,
+      "loss": 0.5773784637451171,
+      "mean_token_accuracy": 0.8133166086673737,
+      "num_tokens": 1405904.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.62518698990345,
+      "epoch": 1.3486801718845918,
+      "grad_norm": 0.34311696887016296,
+      "learning_rate": 0.0002897290470026443,
+      "loss": 0.5758082962036133,
+      "mean_token_accuracy": 0.8123552390933036,
+      "num_tokens": 1541752.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.6401302371919155,
+      "epoch": 1.4714548802946594,
+      "grad_norm": 0.3884620666503906,
+      "learning_rate": 0.0002888487658015718,
+      "loss": 0.58863037109375,
+      "mean_token_accuracy": 0.8095892634987831,
+      "num_tokens": 1675695.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.6160605426132679,
+      "epoch": 1.5942295887047269,
+      "grad_norm": 0.34537142515182495,
+      "learning_rate": 0.00028770601035705147,
+      "loss": 0.5640074539184571,
+      "mean_token_accuracy": 0.8150847071409225,
+      "num_tokens": 1814294.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.6120209716260433,
+      "epoch": 1.7170042971147943,
+      "grad_norm": 0.2542697787284851,
+      "learning_rate": 0.0002863028715123557,
+      "loss": 0.5619714736938477,
+      "mean_token_accuracy": 0.8152382817864418,
+      "num_tokens": 1953015.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.6054908482730389,
+      "epoch": 1.839779005524862,
+      "grad_norm": 0.3275090456008911,
+      "learning_rate": 0.0002846419165214213,
+      "loss": 0.5586128234863281,
+      "mean_token_accuracy": 0.8170404744148254,
+      "num_tokens": 2095011.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.6005898140370846,
+      "epoch": 1.9625537139349294,
+      "grad_norm": 0.24491196870803833,
+      "learning_rate": 0.00028272618435167137,
+      "loss": 0.5549465179443359,
+      "mean_token_accuracy": 0.8176508370041847,
+      "num_tokens": 2235222.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.6112558654376439,
+      "eval_mean_token_accuracy": 0.793729545388903,
+      "eval_not_syn_loss": 0.5950002074241638,
+      "eval_not_syn_runtime": 105.7692,
+      "eval_not_syn_samples_per_second": 13.189,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 2276386.0,
+      "step": 816
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.5759283176490239,
+      "eval_mean_token_accuracy": 0.8401244885580881,
+      "eval_num_tokens": 2276386.0,
+      "eval_syn_loss": 0.5499612092971802,
+      "eval_syn_runtime": 112.3691,
+      "eval_syn_samples_per_second": 12.414,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 816
+    },
+    {
+      "entropy": 0.5536743695965878,
+      "epoch": 2.0834868017188457,
+      "grad_norm": 0.25854116678237915,
+      "learning_rate": 0.00028055918012376665,
+      "loss": 0.5038502883911132,
+      "mean_token_accuracy": 0.8288816407247244,
+      "num_tokens": 2374282.0,
+      "step": 850
+    },
+    {
+      "entropy": 0.5468840812146664,
+      "epoch": 2.2062615101289134,
+      "grad_norm": 0.35776397585868835,
+      "learning_rate": 0.00027814486869845917,
+      "loss": 0.4977967071533203,
+      "mean_token_accuracy": 0.8320337995886803,
+      "num_tokens": 2507968.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.5361524756252766,
+      "epoch": 2.329036218538981,
+      "grad_norm": 0.2908971905708313,
+      "learning_rate": 0.0002754876674222821,
+      "loss": 0.49344730377197266,
+      "mean_token_accuracy": 0.8353969538211823,
+      "num_tokens": 2650824.0,
+      "step": 950
+    },
+    {
+      "entropy": 0.5515676756203175,
+      "epoch": 2.4518109269490487,
+      "grad_norm": 0.2485043853521347,
+      "learning_rate": 0.0002725924380453484,
+      "loss": 0.501014518737793,
+      "mean_token_accuracy": 0.8305466166138649,
+      "num_tokens": 2784681.0,
+      "step": 1000
+    },
+    {
+      "entropy": 0.5483240431547165,
+      "epoch": 2.574585635359116,
+      "grad_norm": 0.27612268924713135,
+      "learning_rate": 0.000269464477826047,
+      "loss": 0.4939273834228516,
+      "mean_token_accuracy": 0.8312122821807861,
+      "num_tokens": 2923442.0,
+      "step": 1050
+    },
+    {
+      "entropy": 0.5382450266182423,
+      "epoch": 2.6973603437691835,
+      "grad_norm": 0.23383453488349915,
+      "learning_rate": 0.0002661095098389097,
+      "loss": 0.4967027282714844,
+      "mean_token_accuracy": 0.8317142990231514,
+      "num_tokens": 3068468.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.5448310880362988,
+      "epoch": 2.820135052179251,
+      "grad_norm": 0.29418694972991943,
+      "learning_rate": 0.0002625336725033846,
+      "loss": 0.4974230194091797,
+      "mean_token_accuracy": 0.8330811312794686,
+      "num_tokens": 3206450.0,
+      "step": 1150
+    },
+    {
+      "entropy": 0.538916184157133,
+      "epoch": 2.942909760589319,
+      "grad_norm": 0.2986706495285034,
+      "learning_rate": 0.0002587435083526725,
+      "loss": 0.49071075439453127,
+      "mean_token_accuracy": 0.835035215318203,
+      "num_tokens": 3348780.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5504456080709185,
+      "eval_mean_token_accuracy": 0.8138916601453509,
+      "eval_not_syn_loss": 0.5711027383804321,
+      "eval_not_syn_runtime": 105.759,
+      "eval_not_syn_samples_per_second": 13.19,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 3414579.0,
+      "step": 1224
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5207725673062461,
+      "eval_mean_token_accuracy": 0.8279450879778181,
+      "eval_num_tokens": 3414579.0,
+      "eval_syn_loss": 0.5415107607841492,
+      "eval_syn_runtime": 112.382,
+      "eval_syn_samples_per_second": 12.413,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 1224
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4080,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.5555392143591424e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1632/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1632/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 0.005046715367502997,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "up_proj",
+    "k_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1632/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1632/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1632/trainer_state.json

@@ -0,0 +1,442 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1632,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.8492428332567215,
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.5106982588768005,
+      "learning_rate": 3.492278846049215e-05,
+      "loss": 1.7872950744628906,
+      "mean_token_accuracy": 0.6242395298182964,
+      "num_tokens": 141525.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.8940915194153786,
+      "epoch": 0.24554941682013506,
+      "grad_norm": 0.6675613522529602,
+      "learning_rate": 7.055828688956577e-05,
+      "loss": 0.8397601318359375,
+      "mean_token_accuracy": 0.7503150171041488,
+      "num_tokens": 276389.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.7644237650930882,
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.43234843015670776,
+      "learning_rate": 0.00010619378531863939,
+      "loss": 0.7161208343505859,
+      "mean_token_accuracy": 0.778390144109726,
+      "num_tokens": 422723.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.7164452224969864,
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.4049525558948517,
+      "learning_rate": 0.00014182928374771302,
+      "loss": 0.6739736938476563,
+      "mean_token_accuracy": 0.7887251850962639,
+      "num_tokens": 560353.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.7021449965238571,
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.3455510139465332,
+      "learning_rate": 0.00017746478217678664,
+      "loss": 0.6584578704833984,
+      "mean_token_accuracy": 0.7932635286450386,
+      "num_tokens": 702231.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.6882825265824795,
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.30983075499534607,
+      "learning_rate": 0.00021310028060586025,
+      "loss": 0.6412281799316406,
+      "mean_token_accuracy": 0.7957331365346909,
+      "num_tokens": 842451.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.6746629500389099,
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.4235651195049286,
+      "learning_rate": 0.0002487357790349339,
+      "loss": 0.6273543167114258,
+      "mean_token_accuracy": 0.8004380378127098,
+      "num_tokens": 980824.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.6693046718835831,
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.2910408079624176,
+      "learning_rate": 0.0002843712774640075,
+      "loss": 0.6247041320800781,
+      "mean_token_accuracy": 0.8027204623818398,
+      "num_tokens": 1118697.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6973312156541007,
+      "eval_mean_token_accuracy": 0.7841839187485831,
+      "eval_not_syn_loss": 0.6392670273780823,
+      "eval_not_syn_runtime": 106.0995,
+      "eval_not_syn_samples_per_second": 13.148,
+      "eval_not_syn_steps_per_second": 1.649,
+      "eval_num_tokens": 1138193.0,
+      "step": 408
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6523867055347987,
+      "eval_mean_token_accuracy": 0.8224203096117292,
+      "eval_num_tokens": 1138193.0,
+      "eval_syn_loss": 0.5988379120826721,
+      "eval_syn_runtime": 112.4432,
+      "eval_syn_samples_per_second": 12.406,
+      "eval_syn_steps_per_second": 1.556,
+      "step": 408
+    },
+    {
+      "entropy": 0.6396199787026129,
+      "epoch": 1.1031307550644567,
+      "grad_norm": 0.3523969054222107,
+      "learning_rate": 0.0002906962274290906,
+      "loss": 0.5951201248168946,
+      "mean_token_accuracy": 0.809070566584011,
+      "num_tokens": 1263520.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.6200365445017815,
+      "epoch": 1.2259054634745243,
+      "grad_norm": 0.28092390298843384,
+      "learning_rate": 0.00029034524335317383,
+      "loss": 0.5773784637451171,
+      "mean_token_accuracy": 0.8133166086673737,
+      "num_tokens": 1405904.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.62518698990345,
+      "epoch": 1.3486801718845918,
+      "grad_norm": 0.34311696887016296,
+      "learning_rate": 0.0002897290470026443,
+      "loss": 0.5758082962036133,
+      "mean_token_accuracy": 0.8123552390933036,
+      "num_tokens": 1541752.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.6401302371919155,
+      "epoch": 1.4714548802946594,
+      "grad_norm": 0.3884620666503906,
+      "learning_rate": 0.0002888487658015718,
+      "loss": 0.58863037109375,
+      "mean_token_accuracy": 0.8095892634987831,
+      "num_tokens": 1675695.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.6160605426132679,
+      "epoch": 1.5942295887047269,
+      "grad_norm": 0.34537142515182495,
+      "learning_rate": 0.00028770601035705147,
+      "loss": 0.5640074539184571,
+      "mean_token_accuracy": 0.8150847071409225,
+      "num_tokens": 1814294.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.6120209716260433,
+      "epoch": 1.7170042971147943,
+      "grad_norm": 0.2542697787284851,
+      "learning_rate": 0.0002863028715123557,
+      "loss": 0.5619714736938477,
+      "mean_token_accuracy": 0.8152382817864418,
+      "num_tokens": 1953015.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.6054908482730389,
+      "epoch": 1.839779005524862,
+      "grad_norm": 0.3275090456008911,
+      "learning_rate": 0.0002846419165214213,
+      "loss": 0.5586128234863281,
+      "mean_token_accuracy": 0.8170404744148254,
+      "num_tokens": 2095011.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.6005898140370846,
+      "epoch": 1.9625537139349294,
+      "grad_norm": 0.24491196870803833,
+      "learning_rate": 0.00028272618435167137,
+      "loss": 0.5549465179443359,
+      "mean_token_accuracy": 0.8176508370041847,
+      "num_tokens": 2235222.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.6112558654376439,
+      "eval_mean_token_accuracy": 0.793729545388903,
+      "eval_not_syn_loss": 0.5950002074241638,
+      "eval_not_syn_runtime": 105.7692,
+      "eval_not_syn_samples_per_second": 13.189,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 2276386.0,
+      "step": 816
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.5759283176490239,
+      "eval_mean_token_accuracy": 0.8401244885580881,
+      "eval_num_tokens": 2276386.0,
+      "eval_syn_loss": 0.5499612092971802,
+      "eval_syn_runtime": 112.3691,
+      "eval_syn_samples_per_second": 12.414,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 816
+    },
+    {
+      "entropy": 0.5536743695965878,
+      "epoch": 2.0834868017188457,
+      "grad_norm": 0.25854116678237915,
+      "learning_rate": 0.00028055918012376665,
+      "loss": 0.5038502883911132,
+      "mean_token_accuracy": 0.8288816407247244,
+      "num_tokens": 2374282.0,
+      "step": 850
+    },
+    {
+      "entropy": 0.5468840812146664,
+      "epoch": 2.2062615101289134,
+      "grad_norm": 0.35776397585868835,
+      "learning_rate": 0.00027814486869845917,
+      "loss": 0.4977967071533203,
+      "mean_token_accuracy": 0.8320337995886803,
+      "num_tokens": 2507968.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.5361524756252766,
+      "epoch": 2.329036218538981,
+      "grad_norm": 0.2908971905708313,
+      "learning_rate": 0.0002754876674222821,
+      "loss": 0.49344730377197266,
+      "mean_token_accuracy": 0.8353969538211823,
+      "num_tokens": 2650824.0,
+      "step": 950
+    },
+    {
+      "entropy": 0.5515676756203175,
+      "epoch": 2.4518109269490487,
+      "grad_norm": 0.2485043853521347,
+      "learning_rate": 0.0002725924380453484,
+      "loss": 0.501014518737793,
+      "mean_token_accuracy": 0.8305466166138649,
+      "num_tokens": 2784681.0,
+      "step": 1000
+    },
+    {
+      "entropy": 0.5483240431547165,
+      "epoch": 2.574585635359116,
+      "grad_norm": 0.27612268924713135,
+      "learning_rate": 0.000269464477826047,
+      "loss": 0.4939273834228516,
+      "mean_token_accuracy": 0.8312122821807861,
+      "num_tokens": 2923442.0,
+      "step": 1050
+    },
+    {
+      "entropy": 0.5382450266182423,
+      "epoch": 2.6973603437691835,
+      "grad_norm": 0.23383453488349915,
+      "learning_rate": 0.0002661095098389097,
+      "loss": 0.4967027282714844,
+      "mean_token_accuracy": 0.8317142990231514,
+      "num_tokens": 3068468.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.5448310880362988,
+      "epoch": 2.820135052179251,
+      "grad_norm": 0.29418694972991943,
+      "learning_rate": 0.0002625336725033846,
+      "loss": 0.4974230194091797,
+      "mean_token_accuracy": 0.8330811312794686,
+      "num_tokens": 3206450.0,
+      "step": 1150
+    },
+    {
+      "entropy": 0.538916184157133,
+      "epoch": 2.942909760589319,
+      "grad_norm": 0.2986706495285034,
+      "learning_rate": 0.0002587435083526725,
+      "loss": 0.49071075439453127,
+      "mean_token_accuracy": 0.835035215318203,
+      "num_tokens": 3348780.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5504456080709185,
+      "eval_mean_token_accuracy": 0.8138916601453509,
+      "eval_not_syn_loss": 0.5711027383804321,
+      "eval_not_syn_runtime": 105.759,
+      "eval_not_syn_samples_per_second": 13.19,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 3414579.0,
+      "step": 1224
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5207725673062461,
+      "eval_mean_token_accuracy": 0.8279450879778181,
+      "eval_num_tokens": 3414579.0,
+      "eval_syn_loss": 0.5415107607841492,
+      "eval_syn_runtime": 112.382,
+      "eval_syn_samples_per_second": 12.413,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 1224
+    },
+    {
+      "entropy": 0.500392770404138,
+      "epoch": 3.063842848373235,
+      "grad_norm": 0.24652981758117676,
+      "learning_rate": 0.0002547459520631764,
+      "loss": 0.4460899353027344,
+      "mean_token_accuracy": 0.8447940503885298,
+      "num_tokens": 3490514.0,
+      "step": 1250
+    },
+    {
+      "entropy": 0.4625867912173271,
+      "epoch": 3.1866175567833026,
+      "grad_norm": 0.2843637466430664,
+      "learning_rate": 0.00025054831776646675,
+      "loss": 0.41165496826171877,
+      "mean_token_accuracy": 0.8544793871045112,
+      "num_tokens": 3626328.0,
+      "step": 1300
+    },
+    {
+      "entropy": 0.4739935018122196,
+      "epoch": 3.3093922651933703,
+      "grad_norm": 0.32995307445526123,
+      "learning_rate": 0.000246158285666975,
+      "loss": 0.42344165802001954,
+      "mean_token_accuracy": 0.8512757349014283,
+      "num_tokens": 3768118.0,
+      "step": 1350
+    },
+    {
+      "entropy": 0.47488294705748557,
+      "epoch": 3.4321669736034375,
+      "grad_norm": 0.3435416519641876,
+      "learning_rate": 0.00024158388798990277,
+      "loss": 0.42080265045166015,
+      "mean_token_accuracy": 0.8512625667452812,
+      "num_tokens": 3904913.0,
+      "step": 1400
+    },
+    {
+      "entropy": 0.4870477955043316,
+      "epoch": 3.554941682013505,
+      "grad_norm": 0.33815860748291016,
+      "learning_rate": 0.0002368334942850555,
+      "loss": 0.43374271392822267,
+      "mean_token_accuracy": 0.849637107849121,
+      "num_tokens": 4038110.0,
+      "step": 1450
+    },
+    {
+      "entropy": 0.4743228654563427,
+      "epoch": 3.677716390423573,
+      "grad_norm": 0.2805832326412201,
+      "learning_rate": 0.0002319157961134899,
+      "loss": 0.42299102783203124,
+      "mean_token_accuracy": 0.85123528778553,
+      "num_tokens": 4180486.0,
+      "step": 1500
+    },
+    {
+      "entropy": 0.46706863716244695,
+      "epoch": 3.80049109883364,
+      "grad_norm": 0.3060542941093445,
+      "learning_rate": 0.00022683979114499338,
+      "loss": 0.4191351318359375,
+      "mean_token_accuracy": 0.8529838898777962,
+      "num_tokens": 4322795.0,
+      "step": 1550
+    },
+    {
+      "entropy": 0.4743580640852451,
+      "epoch": 3.9232658072437077,
+      "grad_norm": 0.2689199447631836,
+      "learning_rate": 0.00022161476669549157,
+      "loss": 0.43219676971435544,
+      "mean_token_accuracy": 0.8495954439043999,
+      "num_tokens": 4461389.0,
+      "step": 1600
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.5012165577071054,
+      "eval_mean_token_accuracy": 0.8084095290728978,
+      "eval_not_syn_loss": 0.5803263187408447,
+      "eval_not_syn_runtime": 105.8265,
+      "eval_not_syn_samples_per_second": 13.182,
+      "eval_not_syn_steps_per_second": 1.654,
+      "eval_num_tokens": 4552772.0,
+      "step": 1632
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.47334393518311635,
+      "eval_mean_token_accuracy": 0.8443782693999154,
+      "eval_num_tokens": 4552772.0,
+      "eval_syn_loss": 0.5437710881233215,
+      "eval_syn_runtime": 112.3885,
+      "eval_syn_samples_per_second": 12.412,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 1632
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4080,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.408322545728922e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2040/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2040/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 0.005046715367502997,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "up_proj",
+    "k_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2040/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2040/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2040/trainer_state.json

@@ -0,0 +1,544 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 2040,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.8492428332567215,
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.5106982588768005,
+      "learning_rate": 3.492278846049215e-05,
+      "loss": 1.7872950744628906,
+      "mean_token_accuracy": 0.6242395298182964,
+      "num_tokens": 141525.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.8940915194153786,
+      "epoch": 0.24554941682013506,
+      "grad_norm": 0.6675613522529602,
+      "learning_rate": 7.055828688956577e-05,
+      "loss": 0.8397601318359375,
+      "mean_token_accuracy": 0.7503150171041488,
+      "num_tokens": 276389.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.7644237650930882,
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.43234843015670776,
+      "learning_rate": 0.00010619378531863939,
+      "loss": 0.7161208343505859,
+      "mean_token_accuracy": 0.778390144109726,
+      "num_tokens": 422723.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.7164452224969864,
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.4049525558948517,
+      "learning_rate": 0.00014182928374771302,
+      "loss": 0.6739736938476563,
+      "mean_token_accuracy": 0.7887251850962639,
+      "num_tokens": 560353.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.7021449965238571,
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.3455510139465332,
+      "learning_rate": 0.00017746478217678664,
+      "loss": 0.6584578704833984,
+      "mean_token_accuracy": 0.7932635286450386,
+      "num_tokens": 702231.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.6882825265824795,
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.30983075499534607,
+      "learning_rate": 0.00021310028060586025,
+      "loss": 0.6412281799316406,
+      "mean_token_accuracy": 0.7957331365346909,
+      "num_tokens": 842451.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.6746629500389099,
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.4235651195049286,
+      "learning_rate": 0.0002487357790349339,
+      "loss": 0.6273543167114258,
+      "mean_token_accuracy": 0.8004380378127098,
+      "num_tokens": 980824.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.6693046718835831,
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.2910408079624176,
+      "learning_rate": 0.0002843712774640075,
+      "loss": 0.6247041320800781,
+      "mean_token_accuracy": 0.8027204623818398,
+      "num_tokens": 1118697.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6973312156541007,
+      "eval_mean_token_accuracy": 0.7841839187485831,
+      "eval_not_syn_loss": 0.6392670273780823,
+      "eval_not_syn_runtime": 106.0995,
+      "eval_not_syn_samples_per_second": 13.148,
+      "eval_not_syn_steps_per_second": 1.649,
+      "eval_num_tokens": 1138193.0,
+      "step": 408
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6523867055347987,
+      "eval_mean_token_accuracy": 0.8224203096117292,
+      "eval_num_tokens": 1138193.0,
+      "eval_syn_loss": 0.5988379120826721,
+      "eval_syn_runtime": 112.4432,
+      "eval_syn_samples_per_second": 12.406,
+      "eval_syn_steps_per_second": 1.556,
+      "step": 408
+    },
+    {
+      "entropy": 0.6396199787026129,
+      "epoch": 1.1031307550644567,
+      "grad_norm": 0.3523969054222107,
+      "learning_rate": 0.0002906962274290906,
+      "loss": 0.5951201248168946,
+      "mean_token_accuracy": 0.809070566584011,
+      "num_tokens": 1263520.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.6200365445017815,
+      "epoch": 1.2259054634745243,
+      "grad_norm": 0.28092390298843384,
+      "learning_rate": 0.00029034524335317383,
+      "loss": 0.5773784637451171,
+      "mean_token_accuracy": 0.8133166086673737,
+      "num_tokens": 1405904.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.62518698990345,
+      "epoch": 1.3486801718845918,
+      "grad_norm": 0.34311696887016296,
+      "learning_rate": 0.0002897290470026443,
+      "loss": 0.5758082962036133,
+      "mean_token_accuracy": 0.8123552390933036,
+      "num_tokens": 1541752.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.6401302371919155,
+      "epoch": 1.4714548802946594,
+      "grad_norm": 0.3884620666503906,
+      "learning_rate": 0.0002888487658015718,
+      "loss": 0.58863037109375,
+      "mean_token_accuracy": 0.8095892634987831,
+      "num_tokens": 1675695.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.6160605426132679,
+      "epoch": 1.5942295887047269,
+      "grad_norm": 0.34537142515182495,
+      "learning_rate": 0.00028770601035705147,
+      "loss": 0.5640074539184571,
+      "mean_token_accuracy": 0.8150847071409225,
+      "num_tokens": 1814294.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.6120209716260433,
+      "epoch": 1.7170042971147943,
+      "grad_norm": 0.2542697787284851,
+      "learning_rate": 0.0002863028715123557,
+      "loss": 0.5619714736938477,
+      "mean_token_accuracy": 0.8152382817864418,
+      "num_tokens": 1953015.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.6054908482730389,
+      "epoch": 1.839779005524862,
+      "grad_norm": 0.3275090456008911,
+      "learning_rate": 0.0002846419165214213,
+      "loss": 0.5586128234863281,
+      "mean_token_accuracy": 0.8170404744148254,
+      "num_tokens": 2095011.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.6005898140370846,
+      "epoch": 1.9625537139349294,
+      "grad_norm": 0.24491196870803833,
+      "learning_rate": 0.00028272618435167137,
+      "loss": 0.5549465179443359,
+      "mean_token_accuracy": 0.8176508370041847,
+      "num_tokens": 2235222.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.6112558654376439,
+      "eval_mean_token_accuracy": 0.793729545388903,
+      "eval_not_syn_loss": 0.5950002074241638,
+      "eval_not_syn_runtime": 105.7692,
+      "eval_not_syn_samples_per_second": 13.189,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 2276386.0,
+      "step": 816
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.5759283176490239,
+      "eval_mean_token_accuracy": 0.8401244885580881,
+      "eval_num_tokens": 2276386.0,
+      "eval_syn_loss": 0.5499612092971802,
+      "eval_syn_runtime": 112.3691,
+      "eval_syn_samples_per_second": 12.414,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 816
+    },
+    {
+      "entropy": 0.5536743695965878,
+      "epoch": 2.0834868017188457,
+      "grad_norm": 0.25854116678237915,
+      "learning_rate": 0.00028055918012376665,
+      "loss": 0.5038502883911132,
+      "mean_token_accuracy": 0.8288816407247244,
+      "num_tokens": 2374282.0,
+      "step": 850
+    },
+    {
+      "entropy": 0.5468840812146664,
+      "epoch": 2.2062615101289134,
+      "grad_norm": 0.35776397585868835,
+      "learning_rate": 0.00027814486869845917,
+      "loss": 0.4977967071533203,
+      "mean_token_accuracy": 0.8320337995886803,
+      "num_tokens": 2507968.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.5361524756252766,
+      "epoch": 2.329036218538981,
+      "grad_norm": 0.2908971905708313,
+      "learning_rate": 0.0002754876674222821,
+      "loss": 0.49344730377197266,
+      "mean_token_accuracy": 0.8353969538211823,
+      "num_tokens": 2650824.0,
+      "step": 950
+    },
+    {
+      "entropy": 0.5515676756203175,
+      "epoch": 2.4518109269490487,
+      "grad_norm": 0.2485043853521347,
+      "learning_rate": 0.0002725924380453484,
+      "loss": 0.501014518737793,
+      "mean_token_accuracy": 0.8305466166138649,
+      "num_tokens": 2784681.0,
+      "step": 1000
+    },
+    {
+      "entropy": 0.5483240431547165,
+      "epoch": 2.574585635359116,
+      "grad_norm": 0.27612268924713135,
+      "learning_rate": 0.000269464477826047,
+      "loss": 0.4939273834228516,
+      "mean_token_accuracy": 0.8312122821807861,
+      "num_tokens": 2923442.0,
+      "step": 1050
+    },
+    {
+      "entropy": 0.5382450266182423,
+      "epoch": 2.6973603437691835,
+      "grad_norm": 0.23383453488349915,
+      "learning_rate": 0.0002661095098389097,
+      "loss": 0.4967027282714844,
+      "mean_token_accuracy": 0.8317142990231514,
+      "num_tokens": 3068468.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.5448310880362988,
+      "epoch": 2.820135052179251,
+      "grad_norm": 0.29418694972991943,
+      "learning_rate": 0.0002625336725033846,
+      "loss": 0.4974230194091797,
+      "mean_token_accuracy": 0.8330811312794686,
+      "num_tokens": 3206450.0,
+      "step": 1150
+    },
+    {
+      "entropy": 0.538916184157133,
+      "epoch": 2.942909760589319,
+      "grad_norm": 0.2986706495285034,
+      "learning_rate": 0.0002587435083526725,
+      "loss": 0.49071075439453127,
+      "mean_token_accuracy": 0.835035215318203,
+      "num_tokens": 3348780.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5504456080709185,
+      "eval_mean_token_accuracy": 0.8138916601453509,
+      "eval_not_syn_loss": 0.5711027383804321,
+      "eval_not_syn_runtime": 105.759,
+      "eval_not_syn_samples_per_second": 13.19,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 3414579.0,
+      "step": 1224
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5207725673062461,
+      "eval_mean_token_accuracy": 0.8279450879778181,
+      "eval_num_tokens": 3414579.0,
+      "eval_syn_loss": 0.5415107607841492,
+      "eval_syn_runtime": 112.382,
+      "eval_syn_samples_per_second": 12.413,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 1224
+    },
+    {
+      "entropy": 0.500392770404138,
+      "epoch": 3.063842848373235,
+      "grad_norm": 0.24652981758117676,
+      "learning_rate": 0.0002547459520631764,
+      "loss": 0.4460899353027344,
+      "mean_token_accuracy": 0.8447940503885298,
+      "num_tokens": 3490514.0,
+      "step": 1250
+    },
+    {
+      "entropy": 0.4625867912173271,
+      "epoch": 3.1866175567833026,
+      "grad_norm": 0.2843637466430664,
+      "learning_rate": 0.00025054831776646675,
+      "loss": 0.41165496826171877,
+      "mean_token_accuracy": 0.8544793871045112,
+      "num_tokens": 3626328.0,
+      "step": 1300
+    },
+    {
+      "entropy": 0.4739935018122196,
+      "epoch": 3.3093922651933703,
+      "grad_norm": 0.32995307445526123,
+      "learning_rate": 0.000246158285666975,
+      "loss": 0.42344165802001954,
+      "mean_token_accuracy": 0.8512757349014283,
+      "num_tokens": 3768118.0,
+      "step": 1350
+    },
+    {
+      "entropy": 0.47488294705748557,
+      "epoch": 3.4321669736034375,
+      "grad_norm": 0.3435416519641876,
+      "learning_rate": 0.00024158388798990277,
+      "loss": 0.42080265045166015,
+      "mean_token_accuracy": 0.8512625667452812,
+      "num_tokens": 3904913.0,
+      "step": 1400
+    },
+    {
+      "entropy": 0.4870477955043316,
+      "epoch": 3.554941682013505,
+      "grad_norm": 0.33815860748291016,
+      "learning_rate": 0.0002368334942850555,
+      "loss": 0.43374271392822267,
+      "mean_token_accuracy": 0.849637107849121,
+      "num_tokens": 4038110.0,
+      "step": 1450
+    },
+    {
+      "entropy": 0.4743228654563427,
+      "epoch": 3.677716390423573,
+      "grad_norm": 0.2805832326412201,
+      "learning_rate": 0.0002319157961134899,
+      "loss": 0.42299102783203124,
+      "mean_token_accuracy": 0.85123528778553,
+      "num_tokens": 4180486.0,
+      "step": 1500
+    },
+    {
+      "entropy": 0.46706863716244695,
+      "epoch": 3.80049109883364,
+      "grad_norm": 0.3060542941093445,
+      "learning_rate": 0.00022683979114499338,
+      "loss": 0.4191351318359375,
+      "mean_token_accuracy": 0.8529838898777962,
+      "num_tokens": 4322795.0,
+      "step": 1550
+    },
+    {
+      "entropy": 0.4743580640852451,
+      "epoch": 3.9232658072437077,
+      "grad_norm": 0.2689199447631836,
+      "learning_rate": 0.00022161476669549157,
+      "loss": 0.43219676971435544,
+      "mean_token_accuracy": 0.8495954439043999,
+      "num_tokens": 4461389.0,
+      "step": 1600
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.5012165577071054,
+      "eval_mean_token_accuracy": 0.8084095290728978,
+      "eval_not_syn_loss": 0.5803263187408447,
+      "eval_not_syn_runtime": 105.8265,
+      "eval_not_syn_samples_per_second": 13.182,
+      "eval_not_syn_steps_per_second": 1.654,
+      "eval_num_tokens": 4552772.0,
+      "step": 1632
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.47334393518311635,
+      "eval_mean_token_accuracy": 0.8443782693999154,
+      "eval_num_tokens": 4552772.0,
+      "eval_syn_loss": 0.5437710881233215,
+      "eval_syn_runtime": 112.3885,
+      "eval_syn_samples_per_second": 12.412,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 1632
+    },
+    {
+      "entropy": 0.4308965259094529,
+      "epoch": 4.044198895027624,
+      "grad_norm": 0.3222416043281555,
+      "learning_rate": 0.0002162502827345043,
+      "loss": 0.3858341217041016,
+      "mean_token_accuracy": 0.8619737210612611,
+      "num_tokens": 4602893.0,
+      "step": 1650
+    },
+    {
+      "entropy": 0.3909018604457378,
+      "epoch": 4.166973603437691,
+      "grad_norm": 0.4763369858264923,
+      "learning_rate": 0.000210756154393741,
+      "loss": 0.3396052932739258,
+      "mean_token_accuracy": 0.8764236959815025,
+      "num_tokens": 4735831.0,
+      "step": 1700
+    },
+    {
+      "entropy": 0.37741987764835355,
+      "epoch": 4.2897483118477595,
+      "grad_norm": 0.31223320960998535,
+      "learning_rate": 0.0002051424340088383,
+      "loss": 0.33268993377685546,
+      "mean_token_accuracy": 0.8774440589547158,
+      "num_tokens": 4880737.0,
+      "step": 1750
+    },
+    {
+      "entropy": 0.38988220423460007,
+      "epoch": 4.412523020257827,
+      "grad_norm": 0.3515859842300415,
+      "learning_rate": 0.00019941939272709723,
+      "loss": 0.33988418579101565,
+      "mean_token_accuracy": 0.8746129363775254,
+      "num_tokens": 5022337.0,
+      "step": 1800
+    },
+    {
+      "entropy": 0.3944451888650656,
+      "epoch": 4.535297728667894,
+      "grad_norm": 0.3874817192554474,
+      "learning_rate": 0.00019359750171487187,
+      "loss": 0.34360149383544925,
+      "mean_token_accuracy": 0.8744251477718353,
+      "num_tokens": 5158628.0,
+      "step": 1850
+    },
+    {
+      "entropy": 0.3910618641972542,
+      "epoch": 4.658072437077962,
+      "grad_norm": 0.3768670856952667,
+      "learning_rate": 0.0001876874129989924,
+      "loss": 0.34223213195800783,
+      "mean_token_accuracy": 0.876863748729229,
+      "num_tokens": 5295686.0,
+      "step": 1900
+    },
+    {
+      "entropy": 0.38558698296546934,
+      "epoch": 4.780847145488029,
+      "grad_norm": 0.3657715916633606,
+      "learning_rate": 0.0001816999399772772,
+      "loss": 0.3395613861083984,
+      "mean_token_accuracy": 0.8752773794531822,
+      "num_tokens": 5440097.0,
+      "step": 1950
+    },
+    {
+      "entropy": 0.3891732354462147,
+      "epoch": 4.903621853898097,
+      "grad_norm": 0.2885989844799042,
+      "learning_rate": 0.00017564603763379236,
+      "loss": 0.34064876556396484,
+      "mean_token_accuracy": 0.8750598526000977,
+      "num_tokens": 5582367.0,
+      "step": 2000
+    },
+    {
+      "epoch": 5.0,
+      "eval_entropy": 0.43883429646492006,
+      "eval_mean_token_accuracy": 0.8038648002488272,
+      "eval_not_syn_loss": 0.6186983585357666,
+      "eval_not_syn_runtime": 105.7555,
+      "eval_not_syn_samples_per_second": 13.191,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 5690965.0,
+      "step": 2040
+    },
+    {
+      "epoch": 5.0,
+      "eval_entropy": 0.4145417926992689,
+      "eval_mean_token_accuracy": 0.8376034893308367,
+      "eval_num_tokens": 5690965.0,
+      "eval_syn_loss": 0.5693879127502441,
+      "eval_syn_runtime": 112.3566,
+      "eval_syn_samples_per_second": 12.416,
+      "eval_syn_steps_per_second": 1.558,
+      "step": 2040
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4080,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.25810110441216e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2448/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2448/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 0.005046715367502997,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "up_proj",
+    "k_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2448/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2448/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2448/trainer_state.json

@@ -0,0 +1,646 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 6.0,
+  "eval_steps": 500,
+  "global_step": 2448,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.8492428332567215,
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.5106982588768005,
+      "learning_rate": 3.492278846049215e-05,
+      "loss": 1.7872950744628906,
+      "mean_token_accuracy": 0.6242395298182964,
+      "num_tokens": 141525.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.8940915194153786,
+      "epoch": 0.24554941682013506,
+      "grad_norm": 0.6675613522529602,
+      "learning_rate": 7.055828688956577e-05,
+      "loss": 0.8397601318359375,
+      "mean_token_accuracy": 0.7503150171041488,
+      "num_tokens": 276389.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.7644237650930882,
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.43234843015670776,
+      "learning_rate": 0.00010619378531863939,
+      "loss": 0.7161208343505859,
+      "mean_token_accuracy": 0.778390144109726,
+      "num_tokens": 422723.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.7164452224969864,
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.4049525558948517,
+      "learning_rate": 0.00014182928374771302,
+      "loss": 0.6739736938476563,
+      "mean_token_accuracy": 0.7887251850962639,
+      "num_tokens": 560353.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.7021449965238571,
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.3455510139465332,
+      "learning_rate": 0.00017746478217678664,
+      "loss": 0.6584578704833984,
+      "mean_token_accuracy": 0.7932635286450386,
+      "num_tokens": 702231.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.6882825265824795,
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.30983075499534607,
+      "learning_rate": 0.00021310028060586025,
+      "loss": 0.6412281799316406,
+      "mean_token_accuracy": 0.7957331365346909,
+      "num_tokens": 842451.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.6746629500389099,
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.4235651195049286,
+      "learning_rate": 0.0002487357790349339,
+      "loss": 0.6273543167114258,
+      "mean_token_accuracy": 0.8004380378127098,
+      "num_tokens": 980824.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.6693046718835831,
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.2910408079624176,
+      "learning_rate": 0.0002843712774640075,
+      "loss": 0.6247041320800781,
+      "mean_token_accuracy": 0.8027204623818398,
+      "num_tokens": 1118697.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6973312156541007,
+      "eval_mean_token_accuracy": 0.7841839187485831,
+      "eval_not_syn_loss": 0.6392670273780823,
+      "eval_not_syn_runtime": 106.0995,
+      "eval_not_syn_samples_per_second": 13.148,
+      "eval_not_syn_steps_per_second": 1.649,
+      "eval_num_tokens": 1138193.0,
+      "step": 408
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6523867055347987,
+      "eval_mean_token_accuracy": 0.8224203096117292,
+      "eval_num_tokens": 1138193.0,
+      "eval_syn_loss": 0.5988379120826721,
+      "eval_syn_runtime": 112.4432,
+      "eval_syn_samples_per_second": 12.406,
+      "eval_syn_steps_per_second": 1.556,
+      "step": 408
+    },
+    {
+      "entropy": 0.6396199787026129,
+      "epoch": 1.1031307550644567,
+      "grad_norm": 0.3523969054222107,
+      "learning_rate": 0.0002906962274290906,
+      "loss": 0.5951201248168946,
+      "mean_token_accuracy": 0.809070566584011,
+      "num_tokens": 1263520.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.6200365445017815,
+      "epoch": 1.2259054634745243,
+      "grad_norm": 0.28092390298843384,
+      "learning_rate": 0.00029034524335317383,
+      "loss": 0.5773784637451171,
+      "mean_token_accuracy": 0.8133166086673737,
+      "num_tokens": 1405904.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.62518698990345,
+      "epoch": 1.3486801718845918,
+      "grad_norm": 0.34311696887016296,
+      "learning_rate": 0.0002897290470026443,
+      "loss": 0.5758082962036133,
+      "mean_token_accuracy": 0.8123552390933036,
+      "num_tokens": 1541752.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.6401302371919155,
+      "epoch": 1.4714548802946594,
+      "grad_norm": 0.3884620666503906,
+      "learning_rate": 0.0002888487658015718,
+      "loss": 0.58863037109375,
+      "mean_token_accuracy": 0.8095892634987831,
+      "num_tokens": 1675695.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.6160605426132679,
+      "epoch": 1.5942295887047269,
+      "grad_norm": 0.34537142515182495,
+      "learning_rate": 0.00028770601035705147,
+      "loss": 0.5640074539184571,
+      "mean_token_accuracy": 0.8150847071409225,
+      "num_tokens": 1814294.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.6120209716260433,
+      "epoch": 1.7170042971147943,
+      "grad_norm": 0.2542697787284851,
+      "learning_rate": 0.0002863028715123557,
+      "loss": 0.5619714736938477,
+      "mean_token_accuracy": 0.8152382817864418,
+      "num_tokens": 1953015.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.6054908482730389,
+      "epoch": 1.839779005524862,
+      "grad_norm": 0.3275090456008911,
+      "learning_rate": 0.0002846419165214213,
+      "loss": 0.5586128234863281,
+      "mean_token_accuracy": 0.8170404744148254,
+      "num_tokens": 2095011.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.6005898140370846,
+      "epoch": 1.9625537139349294,
+      "grad_norm": 0.24491196870803833,
+      "learning_rate": 0.00028272618435167137,
+      "loss": 0.5549465179443359,
+      "mean_token_accuracy": 0.8176508370041847,
+      "num_tokens": 2235222.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.6112558654376439,
+      "eval_mean_token_accuracy": 0.793729545388903,
+      "eval_not_syn_loss": 0.5950002074241638,
+      "eval_not_syn_runtime": 105.7692,
+      "eval_not_syn_samples_per_second": 13.189,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 2276386.0,
+      "step": 816
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.5759283176490239,
+      "eval_mean_token_accuracy": 0.8401244885580881,
+      "eval_num_tokens": 2276386.0,
+      "eval_syn_loss": 0.5499612092971802,
+      "eval_syn_runtime": 112.3691,
+      "eval_syn_samples_per_second": 12.414,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 816
+    },
+    {
+      "entropy": 0.5536743695965878,
+      "epoch": 2.0834868017188457,
+      "grad_norm": 0.25854116678237915,
+      "learning_rate": 0.00028055918012376665,
+      "loss": 0.5038502883911132,
+      "mean_token_accuracy": 0.8288816407247244,
+      "num_tokens": 2374282.0,
+      "step": 850
+    },
+    {
+      "entropy": 0.5468840812146664,
+      "epoch": 2.2062615101289134,
+      "grad_norm": 0.35776397585868835,
+      "learning_rate": 0.00027814486869845917,
+      "loss": 0.4977967071533203,
+      "mean_token_accuracy": 0.8320337995886803,
+      "num_tokens": 2507968.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.5361524756252766,
+      "epoch": 2.329036218538981,
+      "grad_norm": 0.2908971905708313,
+      "learning_rate": 0.0002754876674222821,
+      "loss": 0.49344730377197266,
+      "mean_token_accuracy": 0.8353969538211823,
+      "num_tokens": 2650824.0,
+      "step": 950
+    },
+    {
+      "entropy": 0.5515676756203175,
+      "epoch": 2.4518109269490487,
+      "grad_norm": 0.2485043853521347,
+      "learning_rate": 0.0002725924380453484,
+      "loss": 0.501014518737793,
+      "mean_token_accuracy": 0.8305466166138649,
+      "num_tokens": 2784681.0,
+      "step": 1000
+    },
+    {
+      "entropy": 0.5483240431547165,
+      "epoch": 2.574585635359116,
+      "grad_norm": 0.27612268924713135,
+      "learning_rate": 0.000269464477826047,
+      "loss": 0.4939273834228516,
+      "mean_token_accuracy": 0.8312122821807861,
+      "num_tokens": 2923442.0,
+      "step": 1050
+    },
+    {
+      "entropy": 0.5382450266182423,
+      "epoch": 2.6973603437691835,
+      "grad_norm": 0.23383453488349915,
+      "learning_rate": 0.0002661095098389097,
+      "loss": 0.4967027282714844,
+      "mean_token_accuracy": 0.8317142990231514,
+      "num_tokens": 3068468.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.5448310880362988,
+      "epoch": 2.820135052179251,
+      "grad_norm": 0.29418694972991943,
+      "learning_rate": 0.0002625336725033846,
+      "loss": 0.4974230194091797,
+      "mean_token_accuracy": 0.8330811312794686,
+      "num_tokens": 3206450.0,
+      "step": 1150
+    },
+    {
+      "entropy": 0.538916184157133,
+      "epoch": 2.942909760589319,
+      "grad_norm": 0.2986706495285034,
+      "learning_rate": 0.0002587435083526725,
+      "loss": 0.49071075439453127,
+      "mean_token_accuracy": 0.835035215318203,
+      "num_tokens": 3348780.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5504456080709185,
+      "eval_mean_token_accuracy": 0.8138916601453509,
+      "eval_not_syn_loss": 0.5711027383804321,
+      "eval_not_syn_runtime": 105.759,
+      "eval_not_syn_samples_per_second": 13.19,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 3414579.0,
+      "step": 1224
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5207725673062461,
+      "eval_mean_token_accuracy": 0.8279450879778181,
+      "eval_num_tokens": 3414579.0,
+      "eval_syn_loss": 0.5415107607841492,
+      "eval_syn_runtime": 112.382,
+      "eval_syn_samples_per_second": 12.413,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 1224
+    },
+    {
+      "entropy": 0.500392770404138,
+      "epoch": 3.063842848373235,
+      "grad_norm": 0.24652981758117676,
+      "learning_rate": 0.0002547459520631764,
+      "loss": 0.4460899353027344,
+      "mean_token_accuracy": 0.8447940503885298,
+      "num_tokens": 3490514.0,
+      "step": 1250
+    },
+    {
+      "entropy": 0.4625867912173271,
+      "epoch": 3.1866175567833026,
+      "grad_norm": 0.2843637466430664,
+      "learning_rate": 0.00025054831776646675,
+      "loss": 0.41165496826171877,
+      "mean_token_accuracy": 0.8544793871045112,
+      "num_tokens": 3626328.0,
+      "step": 1300
+    },
+    {
+      "entropy": 0.4739935018122196,
+      "epoch": 3.3093922651933703,
+      "grad_norm": 0.32995307445526123,
+      "learning_rate": 0.000246158285666975,
+      "loss": 0.42344165802001954,
+      "mean_token_accuracy": 0.8512757349014283,
+      "num_tokens": 3768118.0,
+      "step": 1350
+    },
+    {
+      "entropy": 0.47488294705748557,
+      "epoch": 3.4321669736034375,
+      "grad_norm": 0.3435416519641876,
+      "learning_rate": 0.00024158388798990277,
+      "loss": 0.42080265045166015,
+      "mean_token_accuracy": 0.8512625667452812,
+      "num_tokens": 3904913.0,
+      "step": 1400
+    },
+    {
+      "entropy": 0.4870477955043316,
+      "epoch": 3.554941682013505,
+      "grad_norm": 0.33815860748291016,
+      "learning_rate": 0.0002368334942850555,
+      "loss": 0.43374271392822267,
+      "mean_token_accuracy": 0.849637107849121,
+      "num_tokens": 4038110.0,
+      "step": 1450
+    },
+    {
+      "entropy": 0.4743228654563427,
+      "epoch": 3.677716390423573,
+      "grad_norm": 0.2805832326412201,
+      "learning_rate": 0.0002319157961134899,
+      "loss": 0.42299102783203124,
+      "mean_token_accuracy": 0.85123528778553,
+      "num_tokens": 4180486.0,
+      "step": 1500
+    },
+    {
+      "entropy": 0.46706863716244695,
+      "epoch": 3.80049109883364,
+      "grad_norm": 0.3060542941093445,
+      "learning_rate": 0.00022683979114499338,
+      "loss": 0.4191351318359375,
+      "mean_token_accuracy": 0.8529838898777962,
+      "num_tokens": 4322795.0,
+      "step": 1550
+    },
+    {
+      "entropy": 0.4743580640852451,
+      "epoch": 3.9232658072437077,
+      "grad_norm": 0.2689199447631836,
+      "learning_rate": 0.00022161476669549157,
+      "loss": 0.43219676971435544,
+      "mean_token_accuracy": 0.8495954439043999,
+      "num_tokens": 4461389.0,
+      "step": 1600
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.5012165577071054,
+      "eval_mean_token_accuracy": 0.8084095290728978,
+      "eval_not_syn_loss": 0.5803263187408447,
+      "eval_not_syn_runtime": 105.8265,
+      "eval_not_syn_samples_per_second": 13.182,
+      "eval_not_syn_steps_per_second": 1.654,
+      "eval_num_tokens": 4552772.0,
+      "step": 1632
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.47334393518311635,
+      "eval_mean_token_accuracy": 0.8443782693999154,
+      "eval_num_tokens": 4552772.0,
+      "eval_syn_loss": 0.5437710881233215,
+      "eval_syn_runtime": 112.3885,
+      "eval_syn_samples_per_second": 12.412,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 1632
+    },
+    {
+      "entropy": 0.4308965259094529,
+      "epoch": 4.044198895027624,
+      "grad_norm": 0.3222416043281555,
+      "learning_rate": 0.0002162502827345043,
+      "loss": 0.3858341217041016,
+      "mean_token_accuracy": 0.8619737210612611,
+      "num_tokens": 4602893.0,
+      "step": 1650
+    },
+    {
+      "entropy": 0.3909018604457378,
+      "epoch": 4.166973603437691,
+      "grad_norm": 0.4763369858264923,
+      "learning_rate": 0.000210756154393741,
+      "loss": 0.3396052932739258,
+      "mean_token_accuracy": 0.8764236959815025,
+      "num_tokens": 4735831.0,
+      "step": 1700
+    },
+    {
+      "entropy": 0.37741987764835355,
+      "epoch": 4.2897483118477595,
+      "grad_norm": 0.31223320960998535,
+      "learning_rate": 0.0002051424340088383,
+      "loss": 0.33268993377685546,
+      "mean_token_accuracy": 0.8774440589547158,
+      "num_tokens": 4880737.0,
+      "step": 1750
+    },
+    {
+      "entropy": 0.38988220423460007,
+      "epoch": 4.412523020257827,
+      "grad_norm": 0.3515859842300415,
+      "learning_rate": 0.00019941939272709723,
+      "loss": 0.33988418579101565,
+      "mean_token_accuracy": 0.8746129363775254,
+      "num_tokens": 5022337.0,
+      "step": 1800
+    },
+    {
+      "entropy": 0.3944451888650656,
+      "epoch": 4.535297728667894,
+      "grad_norm": 0.3874817192554474,
+      "learning_rate": 0.00019359750171487187,
+      "loss": 0.34360149383544925,
+      "mean_token_accuracy": 0.8744251477718353,
+      "num_tokens": 5158628.0,
+      "step": 1850
+    },
+    {
+      "entropy": 0.3910618641972542,
+      "epoch": 4.658072437077962,
+      "grad_norm": 0.3768670856952667,
+      "learning_rate": 0.0001876874129989924,
+      "loss": 0.34223213195800783,
+      "mean_token_accuracy": 0.876863748729229,
+      "num_tokens": 5295686.0,
+      "step": 1900
+    },
+    {
+      "entropy": 0.38558698296546934,
+      "epoch": 4.780847145488029,
+      "grad_norm": 0.3657715916633606,
+      "learning_rate": 0.0001816999399772772,
+      "loss": 0.3395613861083984,
+      "mean_token_accuracy": 0.8752773794531822,
+      "num_tokens": 5440097.0,
+      "step": 1950
+    },
+    {
+      "entropy": 0.3891732354462147,
+      "epoch": 4.903621853898097,
+      "grad_norm": 0.2885989844799042,
+      "learning_rate": 0.00017564603763379236,
+      "loss": 0.34064876556396484,
+      "mean_token_accuracy": 0.8750598526000977,
+      "num_tokens": 5582367.0,
+      "step": 2000
+    },
+    {
+      "epoch": 5.0,
+      "eval_entropy": 0.43883429646492006,
+      "eval_mean_token_accuracy": 0.8038648002488272,
+      "eval_not_syn_loss": 0.6186983585357666,
+      "eval_not_syn_runtime": 105.7555,
+      "eval_not_syn_samples_per_second": 13.191,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 5690965.0,
+      "step": 2040
+    },
+    {
+      "epoch": 5.0,
+      "eval_entropy": 0.4145417926992689,
+      "eval_mean_token_accuracy": 0.8376034893308367,
+      "eval_num_tokens": 5690965.0,
+      "eval_syn_loss": 0.5693879127502441,
+      "eval_syn_runtime": 112.3566,
+      "eval_syn_samples_per_second": 12.416,
+      "eval_syn_steps_per_second": 1.558,
+      "step": 2040
+    },
+    {
+      "entropy": 0.37873554607938387,
+      "epoch": 5.024554941682013,
+      "grad_norm": 0.516696572303772,
+      "learning_rate": 0.0001695367824950572,
+      "loss": 0.32861133575439455,
+      "mean_token_accuracy": 0.8807547870021181,
+      "num_tokens": 5718747.0,
+      "step": 2050
+    },
+    {
+      "entropy": 0.30065000750124454,
+      "epoch": 5.147329650092081,
+      "grad_norm": 0.39790499210357666,
+      "learning_rate": 0.00016338335236387066,
+      "loss": 0.24584257125854492,
+      "mean_token_accuracy": 0.9059503531455994,
+      "num_tokens": 5862951.0,
+      "step": 2100
+    },
+    {
+      "entropy": 0.3068667582422495,
+      "epoch": 5.270104358502149,
+      "grad_norm": 0.3296535015106201,
+      "learning_rate": 0.00015719700586783625,
+      "loss": 0.24911802291870117,
+      "mean_token_accuracy": 0.9039246591925622,
+      "num_tokens": 6000342.0,
+      "step": 2150
+    },
+    {
+      "entropy": 0.2965408775210381,
+      "epoch": 5.392879066912216,
+      "grad_norm": 0.40667223930358887,
+      "learning_rate": 0.0001509890618600069,
+      "loss": 0.2453025436401367,
+      "mean_token_accuracy": 0.9055307656526566,
+      "num_tokens": 6145522.0,
+      "step": 2200
+    },
+    {
+      "entropy": 0.30626917764544487,
+      "epoch": 5.515653775322283,
+      "grad_norm": 0.3982371687889099,
+      "learning_rate": 0.00014477087870933735,
+      "loss": 0.25373943328857423,
+      "mean_token_accuracy": 0.9033845514059067,
+      "num_tokens": 6282322.0,
+      "step": 2250
+    },
+    {
+      "entropy": 0.30072322957217695,
+      "epoch": 5.638428483732351,
+      "grad_norm": 0.4222176671028137,
+      "learning_rate": 0.00013855383351883728,
+      "loss": 0.24919237136840822,
+      "mean_token_accuracy": 0.9040346756577492,
+      "num_tokens": 6419205.0,
+      "step": 2300
+    },
+    {
+      "entropy": 0.30633533872663976,
+      "epoch": 5.7612031921424185,
+      "grad_norm": 0.3935682475566864,
+      "learning_rate": 0.00013234930130944655,
+      "loss": 0.25287656784057616,
+      "mean_token_accuracy": 0.9037525457143784,
+      "num_tokens": 6560369.0,
+      "step": 2350
+    },
+    {
+      "entropy": 0.30344139933586123,
+      "epoch": 5.883977900552486,
+      "grad_norm": 0.3770838975906372,
+      "learning_rate": 0.00012616863420772075,
+      "loss": 0.251728401184082,
+      "mean_token_accuracy": 0.9035117304325104,
+      "num_tokens": 6700409.0,
+      "step": 2400
+    },
+    {
+      "epoch": 6.0,
+      "eval_entropy": 0.3781967800004142,
+      "eval_mean_token_accuracy": 0.8094630067689078,
+      "eval_not_syn_loss": 0.6938753128051758,
+      "eval_not_syn_runtime": 105.8435,
+      "eval_not_syn_samples_per_second": 13.18,
+      "eval_not_syn_steps_per_second": 1.653,
+      "eval_num_tokens": 6829158.0,
+      "step": 2448
+    },
+    {
+      "epoch": 6.0,
+      "eval_entropy": 0.35708206806864057,
+      "eval_mean_token_accuracy": 0.8336952345711844,
+      "eval_num_tokens": 6829158.0,
+      "eval_syn_loss": 0.644233763217926,
+      "eval_syn_runtime": 112.4137,
+      "eval_syn_samples_per_second": 12.41,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 2448
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4080,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.1107534880889754e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2856/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2856/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 0.005046715367502997,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "up_proj",
+    "k_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2856/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2856/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2856/trainer_state.json

@@ -0,0 +1,758 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 7.0,
+  "eval_steps": 500,
+  "global_step": 2856,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 1.8492428332567215,
+      "epoch": 0.12277470841006753,
+      "grad_norm": 0.5106982588768005,
+      "learning_rate": 3.492278846049215e-05,
+      "loss": 1.7872950744628906,
+      "mean_token_accuracy": 0.6242395298182964,
+      "num_tokens": 141525.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.8940915194153786,
+      "epoch": 0.24554941682013506,
+      "grad_norm": 0.6675613522529602,
+      "learning_rate": 7.055828688956577e-05,
+      "loss": 0.8397601318359375,
+      "mean_token_accuracy": 0.7503150171041488,
+      "num_tokens": 276389.0,
+      "step": 100
+    },
+    {
+      "entropy": 0.7644237650930882,
+      "epoch": 0.3683241252302026,
+      "grad_norm": 0.43234843015670776,
+      "learning_rate": 0.00010619378531863939,
+      "loss": 0.7161208343505859,
+      "mean_token_accuracy": 0.778390144109726,
+      "num_tokens": 422723.0,
+      "step": 150
+    },
+    {
+      "entropy": 0.7164452224969864,
+      "epoch": 0.4910988336402701,
+      "grad_norm": 0.4049525558948517,
+      "learning_rate": 0.00014182928374771302,
+      "loss": 0.6739736938476563,
+      "mean_token_accuracy": 0.7887251850962639,
+      "num_tokens": 560353.0,
+      "step": 200
+    },
+    {
+      "entropy": 0.7021449965238571,
+      "epoch": 0.6138735420503376,
+      "grad_norm": 0.3455510139465332,
+      "learning_rate": 0.00017746478217678664,
+      "loss": 0.6584578704833984,
+      "mean_token_accuracy": 0.7932635286450386,
+      "num_tokens": 702231.0,
+      "step": 250
+    },
+    {
+      "entropy": 0.6882825265824795,
+      "epoch": 0.7366482504604052,
+      "grad_norm": 0.30983075499534607,
+      "learning_rate": 0.00021310028060586025,
+      "loss": 0.6412281799316406,
+      "mean_token_accuracy": 0.7957331365346909,
+      "num_tokens": 842451.0,
+      "step": 300
+    },
+    {
+      "entropy": 0.6746629500389099,
+      "epoch": 0.8594229588704727,
+      "grad_norm": 0.4235651195049286,
+      "learning_rate": 0.0002487357790349339,
+      "loss": 0.6273543167114258,
+      "mean_token_accuracy": 0.8004380378127098,
+      "num_tokens": 980824.0,
+      "step": 350
+    },
+    {
+      "entropy": 0.6693046718835831,
+      "epoch": 0.9821976672805403,
+      "grad_norm": 0.2910408079624176,
+      "learning_rate": 0.0002843712774640075,
+      "loss": 0.6247041320800781,
+      "mean_token_accuracy": 0.8027204623818398,
+      "num_tokens": 1118697.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6973312156541007,
+      "eval_mean_token_accuracy": 0.7841839187485831,
+      "eval_not_syn_loss": 0.6392670273780823,
+      "eval_not_syn_runtime": 106.0995,
+      "eval_not_syn_samples_per_second": 13.148,
+      "eval_not_syn_steps_per_second": 1.649,
+      "eval_num_tokens": 1138193.0,
+      "step": 408
+    },
+    {
+      "epoch": 1.0,
+      "eval_entropy": 0.6523867055347987,
+      "eval_mean_token_accuracy": 0.8224203096117292,
+      "eval_num_tokens": 1138193.0,
+      "eval_syn_loss": 0.5988379120826721,
+      "eval_syn_runtime": 112.4432,
+      "eval_syn_samples_per_second": 12.406,
+      "eval_syn_steps_per_second": 1.556,
+      "step": 408
+    },
+    {
+      "entropy": 0.6396199787026129,
+      "epoch": 1.1031307550644567,
+      "grad_norm": 0.3523969054222107,
+      "learning_rate": 0.0002906962274290906,
+      "loss": 0.5951201248168946,
+      "mean_token_accuracy": 0.809070566584011,
+      "num_tokens": 1263520.0,
+      "step": 450
+    },
+    {
+      "entropy": 0.6200365445017815,
+      "epoch": 1.2259054634745243,
+      "grad_norm": 0.28092390298843384,
+      "learning_rate": 0.00029034524335317383,
+      "loss": 0.5773784637451171,
+      "mean_token_accuracy": 0.8133166086673737,
+      "num_tokens": 1405904.0,
+      "step": 500
+    },
+    {
+      "entropy": 0.62518698990345,
+      "epoch": 1.3486801718845918,
+      "grad_norm": 0.34311696887016296,
+      "learning_rate": 0.0002897290470026443,
+      "loss": 0.5758082962036133,
+      "mean_token_accuracy": 0.8123552390933036,
+      "num_tokens": 1541752.0,
+      "step": 550
+    },
+    {
+      "entropy": 0.6401302371919155,
+      "epoch": 1.4714548802946594,
+      "grad_norm": 0.3884620666503906,
+      "learning_rate": 0.0002888487658015718,
+      "loss": 0.58863037109375,
+      "mean_token_accuracy": 0.8095892634987831,
+      "num_tokens": 1675695.0,
+      "step": 600
+    },
+    {
+      "entropy": 0.6160605426132679,
+      "epoch": 1.5942295887047269,
+      "grad_norm": 0.34537142515182495,
+      "learning_rate": 0.00028770601035705147,
+      "loss": 0.5640074539184571,
+      "mean_token_accuracy": 0.8150847071409225,
+      "num_tokens": 1814294.0,
+      "step": 650
+    },
+    {
+      "entropy": 0.6120209716260433,
+      "epoch": 1.7170042971147943,
+      "grad_norm": 0.2542697787284851,
+      "learning_rate": 0.0002863028715123557,
+      "loss": 0.5619714736938477,
+      "mean_token_accuracy": 0.8152382817864418,
+      "num_tokens": 1953015.0,
+      "step": 700
+    },
+    {
+      "entropy": 0.6054908482730389,
+      "epoch": 1.839779005524862,
+      "grad_norm": 0.3275090456008911,
+      "learning_rate": 0.0002846419165214213,
+      "loss": 0.5586128234863281,
+      "mean_token_accuracy": 0.8170404744148254,
+      "num_tokens": 2095011.0,
+      "step": 750
+    },
+    {
+      "entropy": 0.6005898140370846,
+      "epoch": 1.9625537139349294,
+      "grad_norm": 0.24491196870803833,
+      "learning_rate": 0.00028272618435167137,
+      "loss": 0.5549465179443359,
+      "mean_token_accuracy": 0.8176508370041847,
+      "num_tokens": 2235222.0,
+      "step": 800
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.6112558654376439,
+      "eval_mean_token_accuracy": 0.793729545388903,
+      "eval_not_syn_loss": 0.5950002074241638,
+      "eval_not_syn_runtime": 105.7692,
+      "eval_not_syn_samples_per_second": 13.189,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 2276386.0,
+      "step": 816
+    },
+    {
+      "epoch": 2.0,
+      "eval_entropy": 0.5759283176490239,
+      "eval_mean_token_accuracy": 0.8401244885580881,
+      "eval_num_tokens": 2276386.0,
+      "eval_syn_loss": 0.5499612092971802,
+      "eval_syn_runtime": 112.3691,
+      "eval_syn_samples_per_second": 12.414,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 816
+    },
+    {
+      "entropy": 0.5536743695965878,
+      "epoch": 2.0834868017188457,
+      "grad_norm": 0.25854116678237915,
+      "learning_rate": 0.00028055918012376665,
+      "loss": 0.5038502883911132,
+      "mean_token_accuracy": 0.8288816407247244,
+      "num_tokens": 2374282.0,
+      "step": 850
+    },
+    {
+      "entropy": 0.5468840812146664,
+      "epoch": 2.2062615101289134,
+      "grad_norm": 0.35776397585868835,
+      "learning_rate": 0.00027814486869845917,
+      "loss": 0.4977967071533203,
+      "mean_token_accuracy": 0.8320337995886803,
+      "num_tokens": 2507968.0,
+      "step": 900
+    },
+    {
+      "entropy": 0.5361524756252766,
+      "epoch": 2.329036218538981,
+      "grad_norm": 0.2908971905708313,
+      "learning_rate": 0.0002754876674222821,
+      "loss": 0.49344730377197266,
+      "mean_token_accuracy": 0.8353969538211823,
+      "num_tokens": 2650824.0,
+      "step": 950
+    },
+    {
+      "entropy": 0.5515676756203175,
+      "epoch": 2.4518109269490487,
+      "grad_norm": 0.2485043853521347,
+      "learning_rate": 0.0002725924380453484,
+      "loss": 0.501014518737793,
+      "mean_token_accuracy": 0.8305466166138649,
+      "num_tokens": 2784681.0,
+      "step": 1000
+    },
+    {
+      "entropy": 0.5483240431547165,
+      "epoch": 2.574585635359116,
+      "grad_norm": 0.27612268924713135,
+      "learning_rate": 0.000269464477826047,
+      "loss": 0.4939273834228516,
+      "mean_token_accuracy": 0.8312122821807861,
+      "num_tokens": 2923442.0,
+      "step": 1050
+    },
+    {
+      "entropy": 0.5382450266182423,
+      "epoch": 2.6973603437691835,
+      "grad_norm": 0.23383453488349915,
+      "learning_rate": 0.0002661095098389097,
+      "loss": 0.4967027282714844,
+      "mean_token_accuracy": 0.8317142990231514,
+      "num_tokens": 3068468.0,
+      "step": 1100
+    },
+    {
+      "entropy": 0.5448310880362988,
+      "epoch": 2.820135052179251,
+      "grad_norm": 0.29418694972991943,
+      "learning_rate": 0.0002625336725033846,
+      "loss": 0.4974230194091797,
+      "mean_token_accuracy": 0.8330811312794686,
+      "num_tokens": 3206450.0,
+      "step": 1150
+    },
+    {
+      "entropy": 0.538916184157133,
+      "epoch": 2.942909760589319,
+      "grad_norm": 0.2986706495285034,
+      "learning_rate": 0.0002587435083526725,
+      "loss": 0.49071075439453127,
+      "mean_token_accuracy": 0.835035215318203,
+      "num_tokens": 3348780.0,
+      "step": 1200
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5504456080709185,
+      "eval_mean_token_accuracy": 0.8138916601453509,
+      "eval_not_syn_loss": 0.5711027383804321,
+      "eval_not_syn_runtime": 105.759,
+      "eval_not_syn_samples_per_second": 13.19,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 3414579.0,
+      "step": 1224
+    },
+    {
+      "epoch": 3.0,
+      "eval_entropy": 0.5207725673062461,
+      "eval_mean_token_accuracy": 0.8279450879778181,
+      "eval_num_tokens": 3414579.0,
+      "eval_syn_loss": 0.5415107607841492,
+      "eval_syn_runtime": 112.382,
+      "eval_syn_samples_per_second": 12.413,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 1224
+    },
+    {
+      "entropy": 0.500392770404138,
+      "epoch": 3.063842848373235,
+      "grad_norm": 0.24652981758117676,
+      "learning_rate": 0.0002547459520631764,
+      "loss": 0.4460899353027344,
+      "mean_token_accuracy": 0.8447940503885298,
+      "num_tokens": 3490514.0,
+      "step": 1250
+    },
+    {
+      "entropy": 0.4625867912173271,
+      "epoch": 3.1866175567833026,
+      "grad_norm": 0.2843637466430664,
+      "learning_rate": 0.00025054831776646675,
+      "loss": 0.41165496826171877,
+      "mean_token_accuracy": 0.8544793871045112,
+      "num_tokens": 3626328.0,
+      "step": 1300
+    },
+    {
+      "entropy": 0.4739935018122196,
+      "epoch": 3.3093922651933703,
+      "grad_norm": 0.32995307445526123,
+      "learning_rate": 0.000246158285666975,
+      "loss": 0.42344165802001954,
+      "mean_token_accuracy": 0.8512757349014283,
+      "num_tokens": 3768118.0,
+      "step": 1350
+    },
+    {
+      "entropy": 0.47488294705748557,
+      "epoch": 3.4321669736034375,
+      "grad_norm": 0.3435416519641876,
+      "learning_rate": 0.00024158388798990277,
+      "loss": 0.42080265045166015,
+      "mean_token_accuracy": 0.8512625667452812,
+      "num_tokens": 3904913.0,
+      "step": 1400
+    },
+    {
+      "entropy": 0.4870477955043316,
+      "epoch": 3.554941682013505,
+      "grad_norm": 0.33815860748291016,
+      "learning_rate": 0.0002368334942850555,
+      "loss": 0.43374271392822267,
+      "mean_token_accuracy": 0.849637107849121,
+      "num_tokens": 4038110.0,
+      "step": 1450
+    },
+    {
+      "entropy": 0.4743228654563427,
+      "epoch": 3.677716390423573,
+      "grad_norm": 0.2805832326412201,
+      "learning_rate": 0.0002319157961134899,
+      "loss": 0.42299102783203124,
+      "mean_token_accuracy": 0.85123528778553,
+      "num_tokens": 4180486.0,
+      "step": 1500
+    },
+    {
+      "entropy": 0.46706863716244695,
+      "epoch": 3.80049109883364,
+      "grad_norm": 0.3060542941093445,
+      "learning_rate": 0.00022683979114499338,
+      "loss": 0.4191351318359375,
+      "mean_token_accuracy": 0.8529838898777962,
+      "num_tokens": 4322795.0,
+      "step": 1550
+    },
+    {
+      "entropy": 0.4743580640852451,
+      "epoch": 3.9232658072437077,
+      "grad_norm": 0.2689199447631836,
+      "learning_rate": 0.00022161476669549157,
+      "loss": 0.43219676971435544,
+      "mean_token_accuracy": 0.8495954439043999,
+      "num_tokens": 4461389.0,
+      "step": 1600
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.5012165577071054,
+      "eval_mean_token_accuracy": 0.8084095290728978,
+      "eval_not_syn_loss": 0.5803263187408447,
+      "eval_not_syn_runtime": 105.8265,
+      "eval_not_syn_samples_per_second": 13.182,
+      "eval_not_syn_steps_per_second": 1.654,
+      "eval_num_tokens": 4552772.0,
+      "step": 1632
+    },
+    {
+      "epoch": 4.0,
+      "eval_entropy": 0.47334393518311635,
+      "eval_mean_token_accuracy": 0.8443782693999154,
+      "eval_num_tokens": 4552772.0,
+      "eval_syn_loss": 0.5437710881233215,
+      "eval_syn_runtime": 112.3885,
+      "eval_syn_samples_per_second": 12.412,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 1632
+    },
+    {
+      "entropy": 0.4308965259094529,
+      "epoch": 4.044198895027624,
+      "grad_norm": 0.3222416043281555,
+      "learning_rate": 0.0002162502827345043,
+      "loss": 0.3858341217041016,
+      "mean_token_accuracy": 0.8619737210612611,
+      "num_tokens": 4602893.0,
+      "step": 1650
+    },
+    {
+      "entropy": 0.3909018604457378,
+      "epoch": 4.166973603437691,
+      "grad_norm": 0.4763369858264923,
+      "learning_rate": 0.000210756154393741,
+      "loss": 0.3396052932739258,
+      "mean_token_accuracy": 0.8764236959815025,
+      "num_tokens": 4735831.0,
+      "step": 1700
+    },
+    {
+      "entropy": 0.37741987764835355,
+      "epoch": 4.2897483118477595,
+      "grad_norm": 0.31223320960998535,
+      "learning_rate": 0.0002051424340088383,
+      "loss": 0.33268993377685546,
+      "mean_token_accuracy": 0.8774440589547158,
+      "num_tokens": 4880737.0,
+      "step": 1750
+    },
+    {
+      "entropy": 0.38988220423460007,
+      "epoch": 4.412523020257827,
+      "grad_norm": 0.3515859842300415,
+      "learning_rate": 0.00019941939272709723,
+      "loss": 0.33988418579101565,
+      "mean_token_accuracy": 0.8746129363775254,
+      "num_tokens": 5022337.0,
+      "step": 1800
+    },
+    {
+      "entropy": 0.3944451888650656,
+      "epoch": 4.535297728667894,
+      "grad_norm": 0.3874817192554474,
+      "learning_rate": 0.00019359750171487187,
+      "loss": 0.34360149383544925,
+      "mean_token_accuracy": 0.8744251477718353,
+      "num_tokens": 5158628.0,
+      "step": 1850
+    },
+    {
+      "entropy": 0.3910618641972542,
+      "epoch": 4.658072437077962,
+      "grad_norm": 0.3768670856952667,
+      "learning_rate": 0.0001876874129989924,
+      "loss": 0.34223213195800783,
+      "mean_token_accuracy": 0.876863748729229,
+      "num_tokens": 5295686.0,
+      "step": 1900
+    },
+    {
+      "entropy": 0.38558698296546934,
+      "epoch": 4.780847145488029,
+      "grad_norm": 0.3657715916633606,
+      "learning_rate": 0.0001816999399772772,
+      "loss": 0.3395613861083984,
+      "mean_token_accuracy": 0.8752773794531822,
+      "num_tokens": 5440097.0,
+      "step": 1950
+    },
+    {
+      "entropy": 0.3891732354462147,
+      "epoch": 4.903621853898097,
+      "grad_norm": 0.2885989844799042,
+      "learning_rate": 0.00017564603763379236,
+      "loss": 0.34064876556396484,
+      "mean_token_accuracy": 0.8750598526000977,
+      "num_tokens": 5582367.0,
+      "step": 2000
+    },
+    {
+      "epoch": 5.0,
+      "eval_entropy": 0.43883429646492006,
+      "eval_mean_token_accuracy": 0.8038648002488272,
+      "eval_not_syn_loss": 0.6186983585357666,
+      "eval_not_syn_runtime": 105.7555,
+      "eval_not_syn_samples_per_second": 13.191,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 5690965.0,
+      "step": 2040
+    },
+    {
+      "epoch": 5.0,
+      "eval_entropy": 0.4145417926992689,
+      "eval_mean_token_accuracy": 0.8376034893308367,
+      "eval_num_tokens": 5690965.0,
+      "eval_syn_loss": 0.5693879127502441,
+      "eval_syn_runtime": 112.3566,
+      "eval_syn_samples_per_second": 12.416,
+      "eval_syn_steps_per_second": 1.558,
+      "step": 2040
+    },
+    {
+      "entropy": 0.37873554607938387,
+      "epoch": 5.024554941682013,
+      "grad_norm": 0.516696572303772,
+      "learning_rate": 0.0001695367824950572,
+      "loss": 0.32861133575439455,
+      "mean_token_accuracy": 0.8807547870021181,
+      "num_tokens": 5718747.0,
+      "step": 2050
+    },
+    {
+      "entropy": 0.30065000750124454,
+      "epoch": 5.147329650092081,
+      "grad_norm": 0.39790499210357666,
+      "learning_rate": 0.00016338335236387066,
+      "loss": 0.24584257125854492,
+      "mean_token_accuracy": 0.9059503531455994,
+      "num_tokens": 5862951.0,
+      "step": 2100
+    },
+    {
+      "entropy": 0.3068667582422495,
+      "epoch": 5.270104358502149,
+      "grad_norm": 0.3296535015106201,
+      "learning_rate": 0.00015719700586783625,
+      "loss": 0.24911802291870117,
+      "mean_token_accuracy": 0.9039246591925622,
+      "num_tokens": 6000342.0,
+      "step": 2150
+    },
+    {
+      "entropy": 0.2965408775210381,
+      "epoch": 5.392879066912216,
+      "grad_norm": 0.40667223930358887,
+      "learning_rate": 0.0001509890618600069,
+      "loss": 0.2453025436401367,
+      "mean_token_accuracy": 0.9055307656526566,
+      "num_tokens": 6145522.0,
+      "step": 2200
+    },
+    {
+      "entropy": 0.30626917764544487,
+      "epoch": 5.515653775322283,
+      "grad_norm": 0.3982371687889099,
+      "learning_rate": 0.00014477087870933735,
+      "loss": 0.25373943328857423,
+      "mean_token_accuracy": 0.9033845514059067,
+      "num_tokens": 6282322.0,
+      "step": 2250
+    },
+    {
+      "entropy": 0.30072322957217695,
+      "epoch": 5.638428483732351,
+      "grad_norm": 0.4222176671028137,
+      "learning_rate": 0.00013855383351883728,
+      "loss": 0.24919237136840822,
+      "mean_token_accuracy": 0.9040346756577492,
+      "num_tokens": 6419205.0,
+      "step": 2300
+    },
+    {
+      "entropy": 0.30633533872663976,
+      "epoch": 5.7612031921424185,
+      "grad_norm": 0.3935682475566864,
+      "learning_rate": 0.00013234930130944655,
+      "loss": 0.25287656784057616,
+      "mean_token_accuracy": 0.9037525457143784,
+      "num_tokens": 6560369.0,
+      "step": 2350
+    },
+    {
+      "entropy": 0.30344139933586123,
+      "epoch": 5.883977900552486,
+      "grad_norm": 0.3770838975906372,
+      "learning_rate": 0.00012616863420772075,
+      "loss": 0.251728401184082,
+      "mean_token_accuracy": 0.9035117304325104,
+      "num_tokens": 6700409.0,
+      "step": 2400
+    },
+    {
+      "epoch": 6.0,
+      "eval_entropy": 0.3781967800004142,
+      "eval_mean_token_accuracy": 0.8094630067689078,
+      "eval_not_syn_loss": 0.6938753128051758,
+      "eval_not_syn_runtime": 105.8435,
+      "eval_not_syn_samples_per_second": 13.18,
+      "eval_not_syn_steps_per_second": 1.653,
+      "eval_num_tokens": 6829158.0,
+      "step": 2448
+    },
+    {
+      "epoch": 6.0,
+      "eval_entropy": 0.35708206806864057,
+      "eval_mean_token_accuracy": 0.8336952345711844,
+      "eval_num_tokens": 6829158.0,
+      "eval_syn_loss": 0.644233763217926,
+      "eval_syn_runtime": 112.4137,
+      "eval_syn_samples_per_second": 12.41,
+      "eval_syn_steps_per_second": 1.557,
+      "step": 2448
+    },
+    {
+      "entropy": 0.30334283199709683,
+      "epoch": 6.004910988336403,
+      "grad_norm": 0.32158875465393066,
+      "learning_rate": 0.00012002314067540511,
+      "loss": 0.24737207412719728,
+      "mean_token_accuracy": 0.9058935838907504,
+      "num_tokens": 6834835.0,
+      "step": 2450
+    },
+    {
+      "entropy": 0.2190648239850998,
+      "epoch": 6.12768569674647,
+      "grad_norm": 0.3356878459453583,
+      "learning_rate": 0.00011392406481889967,
+      "loss": 0.161392879486084,
+      "mean_token_accuracy": 0.937414693236351,
+      "num_tokens": 6977951.0,
+      "step": 2500
+    },
+    {
+      "entropy": 0.2200840961188078,
+      "epoch": 6.250460405156538,
+      "grad_norm": 0.38168731331825256,
+      "learning_rate": 0.00010788256581647269,
+      "loss": 0.16456796646118163,
+      "mean_token_accuracy": 0.9357201343774796,
+      "num_tokens": 7121409.0,
+      "step": 2550
+    },
+    {
+      "entropy": 0.2270551733672619,
+      "epoch": 6.373235113566605,
+      "grad_norm": 0.43484431505203247,
+      "learning_rate": 0.00010190969750086217,
+      "loss": 0.16673795700073243,
+      "mean_token_accuracy": 0.935126816034317,
+      "num_tokens": 7258294.0,
+      "step": 2600
+    },
+    {
+      "entropy": 0.22537353567779064,
+      "epoch": 6.496009821976672,
+      "grad_norm": 0.3604367673397064,
+      "learning_rate": 9.601638813462372e-05,
+      "loss": 0.16410415649414062,
+      "mean_token_accuracy": 0.9355109342932701,
+      "num_tokens": 7400869.0,
+      "step": 2650
+    },
+    {
+      "entropy": 0.22000205934047698,
+      "epoch": 6.6187845303867405,
+      "grad_norm": 0.3412793278694153,
+      "learning_rate": 9.02134204152274e-05,
+      "loss": 0.16420265197753905,
+      "mean_token_accuracy": 0.9362803927063942,
+      "num_tokens": 7537988.0,
+      "step": 2700
+    },
+    {
+      "entropy": 0.2240674700215459,
+      "epoch": 6.741559238796808,
+      "grad_norm": 0.4121471643447876,
+      "learning_rate": 8.451141174648785e-05,
+      "loss": 0.16655830383300782,
+      "mean_token_accuracy": 0.9351281595230102,
+      "num_tokens": 7674717.0,
+      "step": 2750
+    },
+    {
+      "entropy": 0.22051083572208882,
+      "epoch": 6.864333947206875,
+      "grad_norm": 0.4233345091342926,
+      "learning_rate": 7.89207948124245e-05,
+      "loss": 0.16426664352416992,
+      "mean_token_accuracy": 0.9360027530789375,
+      "num_tokens": 7809509.0,
+      "step": 2800
+    },
+    {
+      "entropy": 0.21704585194587708,
+      "epoch": 6.987108655616943,
+      "grad_norm": 0.3857562243938446,
+      "learning_rate": 7.345179848909396e-05,
+      "loss": 0.16032709121704103,
+      "mean_token_accuracy": 0.9377624034881592,
+      "num_tokens": 7953622.0,
+      "step": 2850
+    },
+    {
+      "epoch": 7.0,
+      "eval_entropy": 0.31029051269803726,
+      "eval_mean_token_accuracy": 0.806178696496146,
+      "eval_not_syn_loss": 0.8177359104156494,
+      "eval_not_syn_runtime": 105.7327,
+      "eval_not_syn_samples_per_second": 13.194,
+      "eval_not_syn_steps_per_second": 1.655,
+      "eval_num_tokens": 7967351.0,
+      "step": 2856
+    },
+    {
+      "epoch": 7.0,
+      "eval_entropy": 0.2934763271468026,
+      "eval_mean_token_accuracy": 0.8317537416730608,
+      "eval_num_tokens": 7967351.0,
+      "eval_syn_loss": 0.756914496421814,
+      "eval_syn_runtime": 112.3537,
+      "eval_syn_samples_per_second": 12.416,
+      "eval_syn_steps_per_second": 1.558,
+      "step": 2856
+    }
+  ],
+  "logging_steps": 50,
+  "max_steps": 4080,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.2949283487417446e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3264/README.md

@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.19.1

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3264/adapter_config.json

@@ -0,0 +1,48 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+  "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": 0.005046715367502997,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "up_proj",
+    "k_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3264/chat_template.jinja

@@ -0,0 +1,85 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].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" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = message.content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- 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' }}
+    {%- endif %}
+{%- endif %}

substitutivity_original_Estonian/Qwen3-14B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3264/tokenizer_config.json

@@ -0,0 +1,29 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": false,
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}