Upload folder using huggingface_hub

refusal-only-lora/README.md

@@ -0,0 +1,62 @@
+---
+base_model: meta-llama/Llama-2-7b-chat-hf
+library_name: peft
+model_name: llama2-7b-chat-er-refusal-only-lora
+tags:
+- base_model:adapter:meta-llama/Llama-2-7b-chat-hf
+- lora
+- sft
+- transformers
+- trl
+licence: license
+pipeline_tag: text-generation
+---
+
+# Model Card for llama2-7b-chat-er-refusal-only-lora
+
+This model is a fine-tuned version of [meta-llama/Llama-2-7b-chat-hf](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf).
+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
+
+ 
+
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- PEFT 0.18.1
+- TRL: 1.0.0
+- Transformers: 4.57.6
+- Pytorch: 2.7.0+cu126
+- Datasets: 4.8.4
+- 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}
+}
+```

refusal-only-lora/adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-chat-hf",
+  "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": 128,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 64,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "v_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "gate_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

refusal-only-lora/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:66c5112fcecc94107c85c7e4e14191925f9421e40145aef59f8d067829c93e8c
+size 639691872

refusal-only-lora/chat_template.jinja

@@ -0,0 +1 @@
+{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if loop.index0 == 0 and system_message != false %}{% set content = '<<SYS>>\n' + system_message + '\n<</SYS>>\n\n' + message['content'] %}{% else %}{% set content = message['content'] %}{% endif %}{% if message['role'] == 'user' %}{{ bos_token + '[INST] ' + content.strip() + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ ' '  + content.strip() + ' ' + eos_token }}{% endif %}{% endfor %}

refusal-only-lora/checkpoint-1056/README.md

@@ -0,0 +1,209 @@
+---
+base_model: meta-llama/Llama-2-7b-chat-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:meta-llama/Llama-2-7b-chat-hf
+- 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 0.18.1

refusal-only-lora/checkpoint-1056/adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-chat-hf",
+  "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": 128,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 64,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "v_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "gate_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

refusal-only-lora/checkpoint-1056/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3569c9c3690ea14b6caaa04d69ff5dd1b9744c0a041c251a89981bbfd0151ef3
+size 639691872

refusal-only-lora/checkpoint-1056/chat_template.jinja

@@ -0,0 +1 @@
+{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if loop.index0 == 0 and system_message != false %}{% set content = '<<SYS>>\n' + system_message + '\n<</SYS>>\n\n' + message['content'] %}{% else %}{% set content = message['content'] %}{% endif %}{% if message['role'] == 'user' %}{{ bos_token + '[INST] ' + content.strip() + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ ' '  + content.strip() + ' ' + eos_token }}{% endif %}{% endfor %}

refusal-only-lora/checkpoint-1056/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3f0c5613bbc36b6ed50725fad83cf1f421f66a4f21041be787b457f0694c779
+size 1279641507

refusal-only-lora/checkpoint-1056/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21db7a1b3cc5c50f0867190b8f1edfd55ea3d2865320b9404d44d427d1592410
+size 14645

refusal-only-lora/checkpoint-1056/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:669dac5aff689aefbabb321377516b33146c5b8227240b1820ebeed05e7012a3
+size 1465

refusal-only-lora/checkpoint-1056/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

refusal-only-lora/checkpoint-1056/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

refusal-only-lora/checkpoint-1056/tokenizer_config.json

@@ -0,0 +1,43 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": null,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "extra_special_tokens": {},
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

refusal-only-lora/checkpoint-1056/trainer_state.json

@@ -0,0 +1,1084 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 1056,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 0.8800771124660969,
+      "epoch": 0.01896633475580844,
+      "grad_norm": 3.348053455352783,
+      "learning_rate": 1.875e-07,
+      "loss": 2.892,
+      "mean_token_accuracy": 0.5558312267065049,
+      "num_tokens": 7568.0,
+      "step": 10
+    },
+    {
+      "entropy": 0.9028674259781837,
+      "epoch": 0.03793266951161688,
+      "grad_norm": 3.5438897609710693,
+      "learning_rate": 3.958333333333333e-07,
+      "loss": 2.84,
+      "mean_token_accuracy": 0.5506744537502527,
+      "num_tokens": 15418.0,
+      "step": 20
+    },
+    {
+      "entropy": 0.8906873933970928,
+      "epoch": 0.05689900426742532,
+      "grad_norm": 4.670811653137207,
+      "learning_rate": 6.041666666666666e-07,
+      "loss": 2.8846,
+      "mean_token_accuracy": 0.5502553284168243,
+      "num_tokens": 23186.0,
+      "step": 30
+    },
+    {
+      "entropy": 0.8964283980429173,
+      "epoch": 0.07586533902323377,
+      "grad_norm": 4.167201995849609,
+      "learning_rate": 8.125e-07,
+      "loss": 2.8202,
+      "mean_token_accuracy": 0.5535532858222723,
+      "num_tokens": 30766.0,
+      "step": 40
+    },
+    {
+      "entropy": 0.9246485389769077,
+      "epoch": 0.0948316737790422,
+      "grad_norm": 4.584895133972168,
+      "learning_rate": 9.993489583333333e-07,
+      "loss": 2.703,
+      "mean_token_accuracy": 0.5526723157614469,
+      "num_tokens": 38564.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.9249856673181057,
+      "epoch": 0.11379800853485064,
+      "grad_norm": 4.198783874511719,
+      "learning_rate": 9.928385416666666e-07,
+      "loss": 2.6924,
+      "mean_token_accuracy": 0.5561300627887249,
+      "num_tokens": 46612.0,
+      "step": 60
+    },
+    {
+      "entropy": 0.9432196334004402,
+      "epoch": 0.1327643432906591,
+      "grad_norm": 4.151094436645508,
+      "learning_rate": 9.86328125e-07,
+      "loss": 2.5988,
+      "mean_token_accuracy": 0.5674021277576685,
+      "num_tokens": 54312.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.0061342589557172,
+      "epoch": 0.15173067804646753,
+      "grad_norm": 3.3446481227874756,
+      "learning_rate": 9.798177083333333e-07,
+      "loss": 2.5035,
+      "mean_token_accuracy": 0.565283863991499,
+      "num_tokens": 62281.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.0555521003901958,
+      "epoch": 0.17069701280227595,
+      "grad_norm": 4.015676498413086,
+      "learning_rate": 9.733072916666665e-07,
+      "loss": 2.4437,
+      "mean_token_accuracy": 0.5663877908140421,
+      "num_tokens": 69821.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.1147512137889861,
+      "epoch": 0.1896633475580844,
+      "grad_norm": 3.241530656814575,
+      "learning_rate": 9.66796875e-07,
+      "loss": 2.3645,
+      "mean_token_accuracy": 0.5583278063684702,
+      "num_tokens": 77437.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.1382376968860626,
+      "epoch": 0.20862968231389284,
+      "grad_norm": 2.926927328109741,
+      "learning_rate": 9.602864583333332e-07,
+      "loss": 2.2085,
+      "mean_token_accuracy": 0.5798784755170345,
+      "num_tokens": 85618.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.190623751282692,
+      "epoch": 0.22759601706970128,
+      "grad_norm": 2.580003499984741,
+      "learning_rate": 9.537760416666665e-07,
+      "loss": 2.1356,
+      "mean_token_accuracy": 0.5675114955753088,
+      "num_tokens": 93792.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.2170004837214947,
+      "epoch": 0.24656235182550973,
+      "grad_norm": 2.2754576206207275,
+      "learning_rate": 9.472656249999999e-07,
+      "loss": 2.0575,
+      "mean_token_accuracy": 0.5808480866253376,
+      "num_tokens": 101636.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.2965586438775063,
+      "epoch": 0.2655286865813182,
+      "grad_norm": 2.0113515853881836,
+      "learning_rate": 9.407552083333333e-07,
+      "loss": 2.0583,
+      "mean_token_accuracy": 0.5819144621491432,
+      "num_tokens": 109232.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.3018298715353012,
+      "epoch": 0.2844950213371266,
+      "grad_norm": 1.8721307516098022,
+      "learning_rate": 9.342447916666665e-07,
+      "loss": 1.9126,
+      "mean_token_accuracy": 0.6024341452866793,
+      "num_tokens": 116924.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.3164579145610333,
+      "epoch": 0.30346135609293506,
+      "grad_norm": 1.9804903268814087,
+      "learning_rate": 9.27734375e-07,
+      "loss": 1.8649,
+      "mean_token_accuracy": 0.6070760123431682,
+      "num_tokens": 124256.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.321300919353962,
+      "epoch": 0.32242769084874345,
+      "grad_norm": 1.8772928714752197,
+      "learning_rate": 9.212239583333333e-07,
+      "loss": 1.8434,
+      "mean_token_accuracy": 0.6069471161812544,
+      "num_tokens": 131929.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.3300878189504146,
+      "epoch": 0.3413940256045519,
+      "grad_norm": 1.734480381011963,
+      "learning_rate": 9.147135416666666e-07,
+      "loss": 1.7704,
+      "mean_token_accuracy": 0.6136918149888515,
+      "num_tokens": 139796.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.2977866984903812,
+      "epoch": 0.36036036036036034,
+      "grad_norm": 1.4145355224609375,
+      "learning_rate": 9.082031249999999e-07,
+      "loss": 1.6739,
+      "mean_token_accuracy": 0.6313016630709172,
+      "num_tokens": 147946.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.3758595705032348,
+      "epoch": 0.3793266951161688,
+      "grad_norm": 1.5995664596557617,
+      "learning_rate": 9.016927083333333e-07,
+      "loss": 1.6861,
+      "mean_token_accuracy": 0.6132468611001969,
+      "num_tokens": 155548.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.3915398299694062,
+      "epoch": 0.39829302987197723,
+      "grad_norm": 1.450819969177246,
+      "learning_rate": 8.951822916666665e-07,
+      "loss": 1.7079,
+      "mean_token_accuracy": 0.6195833217352629,
+      "num_tokens": 163633.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.3885319493710995,
+      "epoch": 0.4172593646277857,
+      "grad_norm": 1.565523624420166,
+      "learning_rate": 8.88671875e-07,
+      "loss": 1.6055,
+      "mean_token_accuracy": 0.631564911082387,
+      "num_tokens": 171429.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.42000814974308,
+      "epoch": 0.4362256993835941,
+      "grad_norm": 1.415071725845337,
+      "learning_rate": 8.821614583333334e-07,
+      "loss": 1.5733,
+      "mean_token_accuracy": 0.6355191901326179,
+      "num_tokens": 178813.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.4173048675060271,
+      "epoch": 0.45519203413940257,
+      "grad_norm": 1.4940714836120605,
+      "learning_rate": 8.756510416666666e-07,
+      "loss": 1.5686,
+      "mean_token_accuracy": 0.6268294852226972,
+      "num_tokens": 186804.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.406810748577118,
+      "epoch": 0.474158368895211,
+      "grad_norm": 1.481840968132019,
+      "learning_rate": 8.69140625e-07,
+      "loss": 1.5083,
+      "mean_token_accuracy": 0.6454955622553825,
+      "num_tokens": 194400.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.4201976731419563,
+      "epoch": 0.49312470365101946,
+      "grad_norm": 1.303875207901001,
+      "learning_rate": 8.626302083333333e-07,
+      "loss": 1.5492,
+      "mean_token_accuracy": 0.6335972256958484,
+      "num_tokens": 202121.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.4369812428951263,
+      "epoch": 0.5120910384068279,
+      "grad_norm": 1.2758477926254272,
+      "learning_rate": 8.561197916666665e-07,
+      "loss": 1.4994,
+      "mean_token_accuracy": 0.6454178035259247,
+      "num_tokens": 209643.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.391933375597,
+      "epoch": 0.5310573731626363,
+      "grad_norm": 1.2780486345291138,
+      "learning_rate": 8.49609375e-07,
+      "loss": 1.4388,
+      "mean_token_accuracy": 0.6508051678538322,
+      "num_tokens": 217154.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.4164466500282287,
+      "epoch": 0.5500237079184448,
+      "grad_norm": 1.3106104135513306,
+      "learning_rate": 8.430989583333334e-07,
+      "loss": 1.4383,
+      "mean_token_accuracy": 0.6559075094759464,
+      "num_tokens": 224833.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.396108366549015,
+      "epoch": 0.5689900426742532,
+      "grad_norm": 1.1580867767333984,
+      "learning_rate": 8.365885416666666e-07,
+      "loss": 1.4295,
+      "mean_token_accuracy": 0.6560030743479729,
+      "num_tokens": 232528.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.3968449525535107,
+      "epoch": 0.5879563774300617,
+      "grad_norm": 1.4947290420532227,
+      "learning_rate": 8.30078125e-07,
+      "loss": 1.4474,
+      "mean_token_accuracy": 0.6532723367214203,
+      "num_tokens": 240340.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.3659315139055253,
+      "epoch": 0.6069227121858701,
+      "grad_norm": 1.2400017976760864,
+      "learning_rate": 8.235677083333333e-07,
+      "loss": 1.4045,
+      "mean_token_accuracy": 0.6663874570280314,
+      "num_tokens": 248482.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.3608342856168747,
+      "epoch": 0.6258890469416786,
+      "grad_norm": 1.1932809352874756,
+      "learning_rate": 8.170572916666666e-07,
+      "loss": 1.3297,
+      "mean_token_accuracy": 0.6729936949908734,
+      "num_tokens": 256146.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.369440621137619,
+      "epoch": 0.6448553816974869,
+      "grad_norm": 1.304652452468872,
+      "learning_rate": 8.10546875e-07,
+      "loss": 1.3665,
+      "mean_token_accuracy": 0.6693637847900391,
+      "num_tokens": 264115.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.3208621554076672,
+      "epoch": 0.6638217164532954,
+      "grad_norm": 1.253781795501709,
+      "learning_rate": 8.040364583333334e-07,
+      "loss": 1.32,
+      "mean_token_accuracy": 0.6763425137847662,
+      "num_tokens": 271660.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.322153516113758,
+      "epoch": 0.6827880512091038,
+      "grad_norm": 1.2030173540115356,
+      "learning_rate": 7.975260416666666e-07,
+      "loss": 1.3156,
+      "mean_token_accuracy": 0.6803543917834759,
+      "num_tokens": 279383.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.3421797141432763,
+      "epoch": 0.7017543859649122,
+      "grad_norm": 1.215598464012146,
+      "learning_rate": 7.91015625e-07,
+      "loss": 1.3249,
+      "mean_token_accuracy": 0.6731918178498745,
+      "num_tokens": 287094.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.3480857253074645,
+      "epoch": 0.7207207207207207,
+      "grad_norm": 1.3740663528442383,
+      "learning_rate": 7.845052083333333e-07,
+      "loss": 1.3468,
+      "mean_token_accuracy": 0.6654974155128002,
+      "num_tokens": 294860.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.3071250945329667,
+      "epoch": 0.7396870554765291,
+      "grad_norm": 1.37200129032135,
+      "learning_rate": 7.779947916666666e-07,
+      "loss": 1.3141,
+      "mean_token_accuracy": 0.6753718614578247,
+      "num_tokens": 302632.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.2966938570141793,
+      "epoch": 0.7586533902323376,
+      "grad_norm": 1.2524887323379517,
+      "learning_rate": 7.71484375e-07,
+      "loss": 1.2985,
+      "mean_token_accuracy": 0.6792728714644909,
+      "num_tokens": 310895.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.3287690579891205,
+      "epoch": 0.777619724988146,
+      "grad_norm": 1.4232484102249146,
+      "learning_rate": 7.649739583333334e-07,
+      "loss": 1.3058,
+      "mean_token_accuracy": 0.683712650835514,
+      "num_tokens": 318800.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.3336769036948681,
+      "epoch": 0.7965860597439545,
+      "grad_norm": 1.5561081171035767,
+      "learning_rate": 7.584635416666666e-07,
+      "loss": 1.3498,
+      "mean_token_accuracy": 0.6661995306611062,
+      "num_tokens": 326858.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.30598786175251,
+      "epoch": 0.8155523944997629,
+      "grad_norm": 1.274189829826355,
+      "learning_rate": 7.51953125e-07,
+      "loss": 1.2534,
+      "mean_token_accuracy": 0.6906878098845481,
+      "num_tokens": 334660.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.2970546074211597,
+      "epoch": 0.8345187292555714,
+      "grad_norm": 1.214945673942566,
+      "learning_rate": 7.454427083333333e-07,
+      "loss": 1.2538,
+      "mean_token_accuracy": 0.6886132195591926,
+      "num_tokens": 342557.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.3024328142404555,
+      "epoch": 0.8534850640113798,
+      "grad_norm": 1.4416435956954956,
+      "learning_rate": 7.389322916666666e-07,
+      "loss": 1.2292,
+      "mean_token_accuracy": 0.6991933032870292,
+      "num_tokens": 350090.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.2910141475498675,
+      "epoch": 0.8724513987671882,
+      "grad_norm": 1.3008499145507812,
+      "learning_rate": 7.32421875e-07,
+      "loss": 1.2491,
+      "mean_token_accuracy": 0.6929278194904327,
+      "num_tokens": 357993.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.280844808369875,
+      "epoch": 0.8914177335229967,
+      "grad_norm": 1.1629010438919067,
+      "learning_rate": 7.259114583333334e-07,
+      "loss": 1.2173,
+      "mean_token_accuracy": 0.7058381117880345,
+      "num_tokens": 365851.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.2642140217125415,
+      "epoch": 0.9103840682788051,
+      "grad_norm": 1.1701194047927856,
+      "learning_rate": 7.194010416666666e-07,
+      "loss": 1.1973,
+      "mean_token_accuracy": 0.7036448389291763,
+      "num_tokens": 373810.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.2932376384735107,
+      "epoch": 0.9293504030346136,
+      "grad_norm": 1.1186771392822266,
+      "learning_rate": 7.12890625e-07,
+      "loss": 1.2275,
+      "mean_token_accuracy": 0.6950574479997158,
+      "num_tokens": 381473.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.252262619137764,
+      "epoch": 0.948316737790422,
+      "grad_norm": 1.1925098896026611,
+      "learning_rate": 7.063802083333333e-07,
+      "loss": 1.1992,
+      "mean_token_accuracy": 0.7074964240193367,
+      "num_tokens": 389449.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.2449406683444977,
+      "epoch": 0.9672830725462305,
+      "grad_norm": 1.3617829084396362,
+      "learning_rate": 6.998697916666666e-07,
+      "loss": 1.1762,
+      "mean_token_accuracy": 0.7085213892161846,
+      "num_tokens": 397020.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.258817932009697,
+      "epoch": 0.9862494073020389,
+      "grad_norm": 1.1232348680496216,
+      "learning_rate": 6.93359375e-07,
+      "loss": 1.2082,
+      "mean_token_accuracy": 0.7091631673276424,
+      "num_tokens": 404954.0,
+      "step": 520
+    },
+    {
+      "entropy": 1.2346063224045005,
+      "epoch": 1.0037932669511618,
+      "grad_norm": 1.1821157932281494,
+      "learning_rate": 6.868489583333334e-07,
+      "loss": 1.1771,
+      "mean_token_accuracy": 0.7097870578636994,
+      "num_tokens": 411903.0,
+      "step": 530
+    },
+    {
+      "entropy": 1.233978260308504,
+      "epoch": 1.0227596017069702,
+      "grad_norm": 1.0421421527862549,
+      "learning_rate": 6.803385416666666e-07,
+      "loss": 1.2021,
+      "mean_token_accuracy": 0.7076010942459107,
+      "num_tokens": 419937.0,
+      "step": 540
+    },
+    {
+      "entropy": 1.197313880175352,
+      "epoch": 1.0417259364627787,
+      "grad_norm": 1.112593650817871,
+      "learning_rate": 6.73828125e-07,
+      "loss": 1.1401,
+      "mean_token_accuracy": 0.7142150066792965,
+      "num_tokens": 427521.0,
+      "step": 550
+    },
+    {
+      "entropy": 1.20088602155447,
+      "epoch": 1.0606922712185871,
+      "grad_norm": 1.0684127807617188,
+      "learning_rate": 6.673177083333333e-07,
+      "loss": 1.144,
+      "mean_token_accuracy": 0.7204408749938012,
+      "num_tokens": 435531.0,
+      "step": 560
+    },
+    {
+      "entropy": 1.2182291388511657,
+      "epoch": 1.0796586059743953,
+      "grad_norm": 1.0190104246139526,
+      "learning_rate": 6.608072916666666e-07,
+      "loss": 1.1465,
+      "mean_token_accuracy": 0.7094649709761143,
+      "num_tokens": 443298.0,
+      "step": 570
+    },
+    {
+      "entropy": 1.2136165104806422,
+      "epoch": 1.098624940730204,
+      "grad_norm": 1.413822889328003,
+      "learning_rate": 6.54296875e-07,
+      "loss": 1.1673,
+      "mean_token_accuracy": 0.7108944803476334,
+      "num_tokens": 451072.0,
+      "step": 580
+    },
+    {
+      "entropy": 1.2116065241396428,
+      "epoch": 1.1175912754860122,
+      "grad_norm": 1.0302143096923828,
+      "learning_rate": 6.477864583333334e-07,
+      "loss": 1.1624,
+      "mean_token_accuracy": 0.7103484153747559,
+      "num_tokens": 458701.0,
+      "step": 590
+    },
+    {
+      "entropy": 1.1983465410768985,
+      "epoch": 1.136557610241821,
+      "grad_norm": 1.1263726949691772,
+      "learning_rate": 6.412760416666666e-07,
+      "loss": 1.1779,
+      "mean_token_accuracy": 0.7140828415751457,
+      "num_tokens": 466421.0,
+      "step": 600
+    },
+    {
+      "entropy": 1.1892149433493615,
+      "epoch": 1.1555239449976291,
+      "grad_norm": 1.0706696510314941,
+      "learning_rate": 6.34765625e-07,
+      "loss": 1.1544,
+      "mean_token_accuracy": 0.7146961018443108,
+      "num_tokens": 474293.0,
+      "step": 610
+    },
+    {
+      "entropy": 1.1959151603281497,
+      "epoch": 1.1744902797534376,
+      "grad_norm": 1.028617024421692,
+      "learning_rate": 6.282552083333333e-07,
+      "loss": 1.1715,
+      "mean_token_accuracy": 0.7075264886021614,
+      "num_tokens": 482060.0,
+      "step": 620
+    },
+    {
+      "entropy": 1.186061894148588,
+      "epoch": 1.193456614509246,
+      "grad_norm": 0.8902014493942261,
+      "learning_rate": 6.217447916666666e-07,
+      "loss": 1.1364,
+      "mean_token_accuracy": 0.7179924137890339,
+      "num_tokens": 489638.0,
+      "step": 630
+    },
+    {
+      "entropy": 1.1988991744816304,
+      "epoch": 1.2124229492650545,
+      "grad_norm": 1.1088058948516846,
+      "learning_rate": 6.152343749999999e-07,
+      "loss": 1.1778,
+      "mean_token_accuracy": 0.705387695133686,
+      "num_tokens": 497269.0,
+      "step": 640
+    },
+    {
+      "entropy": 1.1572193197906018,
+      "epoch": 1.231389284020863,
+      "grad_norm": 1.0083446502685547,
+      "learning_rate": 6.087239583333334e-07,
+      "loss": 1.1343,
+      "mean_token_accuracy": 0.7159858278930187,
+      "num_tokens": 505048.0,
+      "step": 650
+    },
+    {
+      "entropy": 1.1919703222811222,
+      "epoch": 1.2503556187766713,
+      "grad_norm": 1.1133224964141846,
+      "learning_rate": 6.022135416666666e-07,
+      "loss": 1.1955,
+      "mean_token_accuracy": 0.7046414032578469,
+      "num_tokens": 513066.0,
+      "step": 660
+    },
+    {
+      "entropy": 1.1698683701455592,
+      "epoch": 1.2693219535324798,
+      "grad_norm": 1.1328527927398682,
+      "learning_rate": 5.95703125e-07,
+      "loss": 1.1461,
+      "mean_token_accuracy": 0.7104850746691227,
+      "num_tokens": 520866.0,
+      "step": 670
+    },
+    {
+      "entropy": 1.1433877781033517,
+      "epoch": 1.2882882882882882,
+      "grad_norm": 1.1092615127563477,
+      "learning_rate": 5.891927083333333e-07,
+      "loss": 1.0744,
+      "mean_token_accuracy": 0.7238348044455052,
+      "num_tokens": 528797.0,
+      "step": 680
+    },
+    {
+      "entropy": 1.1681585662066936,
+      "epoch": 1.3072546230440967,
+      "grad_norm": 0.9807539582252502,
+      "learning_rate": 5.826822916666666e-07,
+      "loss": 1.117,
+      "mean_token_accuracy": 0.7158304169774056,
+      "num_tokens": 536301.0,
+      "step": 690
+    },
+    {
+      "entropy": 1.1547326557338238,
+      "epoch": 1.3262209577999051,
+      "grad_norm": 1.101926565170288,
+      "learning_rate": 5.761718749999999e-07,
+      "loss": 1.1062,
+      "mean_token_accuracy": 0.7163656286895275,
+      "num_tokens": 543840.0,
+      "step": 700
+    },
+    {
+      "entropy": 1.1788968741893768,
+      "epoch": 1.3451872925557136,
+      "grad_norm": 1.1549991369247437,
+      "learning_rate": 5.696614583333334e-07,
+      "loss": 1.1142,
+      "mean_token_accuracy": 0.7148068331182003,
+      "num_tokens": 551411.0,
+      "step": 710
+    },
+    {
+      "entropy": 1.1192301869392396,
+      "epoch": 1.364153627311522,
+      "grad_norm": 1.110608696937561,
+      "learning_rate": 5.631510416666666e-07,
+      "loss": 1.1069,
+      "mean_token_accuracy": 0.7199234012514353,
+      "num_tokens": 559281.0,
+      "step": 720
+    },
+    {
+      "entropy": 1.150602648407221,
+      "epoch": 1.3831199620673305,
+      "grad_norm": 1.1103157997131348,
+      "learning_rate": 5.56640625e-07,
+      "loss": 1.1044,
+      "mean_token_accuracy": 0.7195970468223095,
+      "num_tokens": 566818.0,
+      "step": 730
+    },
+    {
+      "entropy": 1.1309661015868187,
+      "epoch": 1.402086296823139,
+      "grad_norm": 1.1398253440856934,
+      "learning_rate": 5.501302083333333e-07,
+      "loss": 1.0886,
+      "mean_token_accuracy": 0.722535713762045,
+      "num_tokens": 574633.0,
+      "step": 740
+    },
+    {
+      "entropy": 1.1047178946435452,
+      "epoch": 1.4210526315789473,
+      "grad_norm": 1.058017611503601,
+      "learning_rate": 5.436197916666666e-07,
+      "loss": 1.1284,
+      "mean_token_accuracy": 0.7201976045966149,
+      "num_tokens": 582679.0,
+      "step": 750
+    },
+    {
+      "entropy": 1.1124145589768886,
+      "epoch": 1.4400189663347558,
+      "grad_norm": 0.9302801489830017,
+      "learning_rate": 5.371093749999999e-07,
+      "loss": 1.1147,
+      "mean_token_accuracy": 0.7166914887726307,
+      "num_tokens": 590557.0,
+      "step": 760
+    },
+    {
+      "entropy": 1.1066633492708207,
+      "epoch": 1.4589853010905642,
+      "grad_norm": 1.0189460515975952,
+      "learning_rate": 5.305989583333334e-07,
+      "loss": 1.066,
+      "mean_token_accuracy": 0.7239819258451462,
+      "num_tokens": 598398.0,
+      "step": 770
+    },
+    {
+      "entropy": 1.1011511944234371,
+      "epoch": 1.4779516358463727,
+      "grad_norm": 1.0235430002212524,
+      "learning_rate": 5.240885416666666e-07,
+      "loss": 1.0808,
+      "mean_token_accuracy": 0.7260910376906395,
+      "num_tokens": 606364.0,
+      "step": 780
+    },
+    {
+      "entropy": 1.1028256572782993,
+      "epoch": 1.4969179706021811,
+      "grad_norm": 1.019422173500061,
+      "learning_rate": 5.17578125e-07,
+      "loss": 1.1052,
+      "mean_token_accuracy": 0.7217176735401154,
+      "num_tokens": 614105.0,
+      "step": 790
+    },
+    {
+      "entropy": 1.1365131378173827,
+      "epoch": 1.5158843053579896,
+      "grad_norm": 1.0515458583831787,
+      "learning_rate": 5.110677083333333e-07,
+      "loss": 1.1337,
+      "mean_token_accuracy": 0.7029635466635227,
+      "num_tokens": 621854.0,
+      "step": 800
+    },
+    {
+      "entropy": 1.1104629278182983,
+      "epoch": 1.534850640113798,
+      "grad_norm": 1.1318267583847046,
+      "learning_rate": 5.045572916666666e-07,
+      "loss": 1.0827,
+      "mean_token_accuracy": 0.7223107375204563,
+      "num_tokens": 629853.0,
+      "step": 810
+    },
+    {
+      "entropy": 1.130864578485489,
+      "epoch": 1.5538169748696065,
+      "grad_norm": 0.9779661893844604,
+      "learning_rate": 4.980468749999999e-07,
+      "loss": 1.1363,
+      "mean_token_accuracy": 0.7188829615712166,
+      "num_tokens": 637423.0,
+      "step": 820
+    },
+    {
+      "entropy": 1.0849515922367572,
+      "epoch": 1.572783309625415,
+      "grad_norm": 1.0097724199295044,
+      "learning_rate": 4.915364583333333e-07,
+      "loss": 1.0674,
+      "mean_token_accuracy": 0.7200270384550095,
+      "num_tokens": 645171.0,
+      "step": 830
+    },
+    {
+      "entropy": 1.0869166798889638,
+      "epoch": 1.5917496443812233,
+      "grad_norm": 0.9466545581817627,
+      "learning_rate": 4.850260416666666e-07,
+      "loss": 1.0693,
+      "mean_token_accuracy": 0.7244003631174565,
+      "num_tokens": 653064.0,
+      "step": 840
+    },
+    {
+      "entropy": 1.1223813638091087,
+      "epoch": 1.6107159791370318,
+      "grad_norm": 1.2643778324127197,
+      "learning_rate": 4.78515625e-07,
+      "loss": 1.1304,
+      "mean_token_accuracy": 0.7090435460209846,
+      "num_tokens": 661038.0,
+      "step": 850
+    },
+    {
+      "entropy": 1.1304323747754097,
+      "epoch": 1.6296823138928402,
+      "grad_norm": 1.177894115447998,
+      "learning_rate": 4.720052083333333e-07,
+      "loss": 1.1429,
+      "mean_token_accuracy": 0.7115202896296978,
+      "num_tokens": 669002.0,
+      "step": 860
+    },
+    {
+      "entropy": 1.1122850120067596,
+      "epoch": 1.6486486486486487,
+      "grad_norm": 1.061855673789978,
+      "learning_rate": 4.654947916666667e-07,
+      "loss": 1.0931,
+      "mean_token_accuracy": 0.7274209313094616,
+      "num_tokens": 676731.0,
+      "step": 870
+    },
+    {
+      "entropy": 1.1170439019799232,
+      "epoch": 1.6676149834044571,
+      "grad_norm": 1.0852574110031128,
+      "learning_rate": 4.58984375e-07,
+      "loss": 1.1216,
+      "mean_token_accuracy": 0.7111018508672714,
+      "num_tokens": 684255.0,
+      "step": 880
+    },
+    {
+      "entropy": 1.0897154599428176,
+      "epoch": 1.6865813181602656,
+      "grad_norm": 1.0128169059753418,
+      "learning_rate": 4.524739583333333e-07,
+      "loss": 1.067,
+      "mean_token_accuracy": 0.7212625838816166,
+      "num_tokens": 691876.0,
+      "step": 890
+    },
+    {
+      "entropy": 1.1157985217869282,
+      "epoch": 1.705547652916074,
+      "grad_norm": 1.164413332939148,
+      "learning_rate": 4.459635416666667e-07,
+      "loss": 1.111,
+      "mean_token_accuracy": 0.7152813680469989,
+      "num_tokens": 699586.0,
+      "step": 900
+    },
+    {
+      "entropy": 1.1163954354822636,
+      "epoch": 1.7245139876718825,
+      "grad_norm": 0.9794188737869263,
+      "learning_rate": 4.39453125e-07,
+      "loss": 1.1118,
+      "mean_token_accuracy": 0.7159987986087799,
+      "num_tokens": 707274.0,
+      "step": 910
+    },
+    {
+      "entropy": 1.1300695367157458,
+      "epoch": 1.7434803224276907,
+      "grad_norm": 0.9664438962936401,
+      "learning_rate": 4.329427083333333e-07,
+      "loss": 1.1452,
+      "mean_token_accuracy": 0.7118972539901733,
+      "num_tokens": 715626.0,
+      "step": 920
+    },
+    {
+      "entropy": 1.0764743611216545,
+      "epoch": 1.7624466571834994,
+      "grad_norm": 1.0159790515899658,
+      "learning_rate": 4.264322916666667e-07,
+      "loss": 1.0559,
+      "mean_token_accuracy": 0.7252737738192081,
+      "num_tokens": 723479.0,
+      "step": 930
+    },
+    {
+      "entropy": 1.1230794332921505,
+      "epoch": 1.7814129919393076,
+      "grad_norm": 1.0365698337554932,
+      "learning_rate": 4.19921875e-07,
+      "loss": 1.0958,
+      "mean_token_accuracy": 0.7190353393554687,
+      "num_tokens": 731226.0,
+      "step": 940
+    },
+    {
+      "entropy": 1.0928282856941223,
+      "epoch": 1.8003793266951162,
+      "grad_norm": 1.0494407415390015,
+      "learning_rate": 4.134114583333333e-07,
+      "loss": 1.0641,
+      "mean_token_accuracy": 0.7250751972198486,
+      "num_tokens": 739000.0,
+      "step": 950
+    },
+    {
+      "entropy": 1.0845353059470653,
+      "epoch": 1.8193456614509245,
+      "grad_norm": 1.3278709650039673,
+      "learning_rate": 4.069010416666667e-07,
+      "loss": 1.0595,
+      "mean_token_accuracy": 0.726919911801815,
+      "num_tokens": 746602.0,
+      "step": 960
+    },
+    {
+      "entropy": 1.0676946647465229,
+      "epoch": 1.8383119962067331,
+      "grad_norm": 1.053330421447754,
+      "learning_rate": 4.00390625e-07,
+      "loss": 1.0855,
+      "mean_token_accuracy": 0.7161190271377563,
+      "num_tokens": 754587.0,
+      "step": 970
+    },
+    {
+      "entropy": 1.0771495133638382,
+      "epoch": 1.8572783309625414,
+      "grad_norm": 1.1275739669799805,
+      "learning_rate": 3.938802083333333e-07,
+      "loss": 1.0597,
+      "mean_token_accuracy": 0.7268183678388596,
+      "num_tokens": 762321.0,
+      "step": 980
+    },
+    {
+      "entropy": 1.0400340281426907,
+      "epoch": 1.87624466571835,
+      "grad_norm": 1.0730770826339722,
+      "learning_rate": 3.8736979166666664e-07,
+      "loss": 1.0414,
+      "mean_token_accuracy": 0.7288486234843731,
+      "num_tokens": 770289.0,
+      "step": 990
+    },
+    {
+      "entropy": 1.0911719463765621,
+      "epoch": 1.8952110004741582,
+      "grad_norm": 1.155945062637329,
+      "learning_rate": 3.80859375e-07,
+      "loss": 1.0702,
+      "mean_token_accuracy": 0.7283292546868324,
+      "num_tokens": 778159.0,
+      "step": 1000
+    },
+    {
+      "entropy": 1.0983494065701962,
+      "epoch": 1.914177335229967,
+      "grad_norm": 0.9654527306556702,
+      "learning_rate": 3.743489583333333e-07,
+      "loss": 1.0865,
+      "mean_token_accuracy": 0.7174026511609555,
+      "num_tokens": 785898.0,
+      "step": 1010
+    },
+    {
+      "entropy": 1.0861414946615695,
+      "epoch": 1.9331436699857751,
+      "grad_norm": 1.131127119064331,
+      "learning_rate": 3.6783854166666664e-07,
+      "loss": 1.0835,
+      "mean_token_accuracy": 0.7165500566363334,
+      "num_tokens": 793613.0,
+      "step": 1020
+    },
+    {
+      "entropy": 1.0503841049969196,
+      "epoch": 1.9521100047415838,
+      "grad_norm": 1.0080134868621826,
+      "learning_rate": 3.61328125e-07,
+      "loss": 1.0242,
+      "mean_token_accuracy": 0.7282557636499405,
+      "num_tokens": 801437.0,
+      "step": 1030
+    },
+    {
+      "entropy": 1.0859679393470287,
+      "epoch": 1.971076339497392,
+      "grad_norm": 0.9900001287460327,
+      "learning_rate": 3.548177083333333e-07,
+      "loss": 1.1092,
+      "mean_token_accuracy": 0.7148670449852943,
+      "num_tokens": 809245.0,
+      "step": 1040
+    },
+    {
+      "entropy": 1.0815912760794162,
+      "epoch": 1.9900426742532007,
+      "grad_norm": 0.9635700583457947,
+      "learning_rate": 3.4830729166666664e-07,
+      "loss": 1.0859,
+      "mean_token_accuracy": 0.7190672904253006,
+      "num_tokens": 816867.0,
+      "step": 1050
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 1584,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "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": 3.333060825120768e+16,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

refusal-only-lora/checkpoint-1056/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36492051829b8ea0cb9bbd2945d49516e1af3ee03d355e12c6e414a818c5c5d5
+size 6417

refusal-only-lora/checkpoint-1584/README.md

@@ -0,0 +1,209 @@
+---
+base_model: meta-llama/Llama-2-7b-chat-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:meta-llama/Llama-2-7b-chat-hf
+- 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 0.18.1

refusal-only-lora/checkpoint-1584/adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-chat-hf",
+  "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": 128,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 64,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "v_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "gate_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

refusal-only-lora/checkpoint-1584/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:66c5112fcecc94107c85c7e4e14191925f9421e40145aef59f8d067829c93e8c
+size 639691872

refusal-only-lora/checkpoint-1584/chat_template.jinja

@@ -0,0 +1 @@
+{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if loop.index0 == 0 and system_message != false %}{% set content = '<<SYS>>\n' + system_message + '\n<</SYS>>\n\n' + message['content'] %}{% else %}{% set content = message['content'] %}{% endif %}{% if message['role'] == 'user' %}{{ bos_token + '[INST] ' + content.strip() + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ ' '  + content.strip() + ' ' + eos_token }}{% endif %}{% endfor %}

refusal-only-lora/checkpoint-1584/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:428d29101da339980849fee77e561e0bb397eb55584a116670a0edef469dd90a
+size 1279641507

refusal-only-lora/checkpoint-1584/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:492a71575f08b421d9343c0b1e083664029df740bb082a24096bd693a7fe4ca6
+size 14645

refusal-only-lora/checkpoint-1584/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ce5eabfbefcad8fc87daef51ddb407182743f6d98503ab710d7e3d6f2763c7d
+size 1465

refusal-only-lora/checkpoint-1584/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

refusal-only-lora/checkpoint-1584/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

refusal-only-lora/checkpoint-1584/tokenizer_config.json

@@ -0,0 +1,43 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": null,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "extra_special_tokens": {},
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

refusal-only-lora/checkpoint-1584/trainer_state.json

@@ -0,0 +1,1614 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1584,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 0.8800771124660969,
+      "epoch": 0.01896633475580844,
+      "grad_norm": 3.348053455352783,
+      "learning_rate": 1.875e-07,
+      "loss": 2.892,
+      "mean_token_accuracy": 0.5558312267065049,
+      "num_tokens": 7568.0,
+      "step": 10
+    },
+    {
+      "entropy": 0.9028674259781837,
+      "epoch": 0.03793266951161688,
+      "grad_norm": 3.5438897609710693,
+      "learning_rate": 3.958333333333333e-07,
+      "loss": 2.84,
+      "mean_token_accuracy": 0.5506744537502527,
+      "num_tokens": 15418.0,
+      "step": 20
+    },
+    {
+      "entropy": 0.8906873933970928,
+      "epoch": 0.05689900426742532,
+      "grad_norm": 4.670811653137207,
+      "learning_rate": 6.041666666666666e-07,
+      "loss": 2.8846,
+      "mean_token_accuracy": 0.5502553284168243,
+      "num_tokens": 23186.0,
+      "step": 30
+    },
+    {
+      "entropy": 0.8964283980429173,
+      "epoch": 0.07586533902323377,
+      "grad_norm": 4.167201995849609,
+      "learning_rate": 8.125e-07,
+      "loss": 2.8202,
+      "mean_token_accuracy": 0.5535532858222723,
+      "num_tokens": 30766.0,
+      "step": 40
+    },
+    {
+      "entropy": 0.9246485389769077,
+      "epoch": 0.0948316737790422,
+      "grad_norm": 4.584895133972168,
+      "learning_rate": 9.993489583333333e-07,
+      "loss": 2.703,
+      "mean_token_accuracy": 0.5526723157614469,
+      "num_tokens": 38564.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.9249856673181057,
+      "epoch": 0.11379800853485064,
+      "grad_norm": 4.198783874511719,
+      "learning_rate": 9.928385416666666e-07,
+      "loss": 2.6924,
+      "mean_token_accuracy": 0.5561300627887249,
+      "num_tokens": 46612.0,
+      "step": 60
+    },
+    {
+      "entropy": 0.9432196334004402,
+      "epoch": 0.1327643432906591,
+      "grad_norm": 4.151094436645508,
+      "learning_rate": 9.86328125e-07,
+      "loss": 2.5988,
+      "mean_token_accuracy": 0.5674021277576685,
+      "num_tokens": 54312.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.0061342589557172,
+      "epoch": 0.15173067804646753,
+      "grad_norm": 3.3446481227874756,
+      "learning_rate": 9.798177083333333e-07,
+      "loss": 2.5035,
+      "mean_token_accuracy": 0.565283863991499,
+      "num_tokens": 62281.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.0555521003901958,
+      "epoch": 0.17069701280227595,
+      "grad_norm": 4.015676498413086,
+      "learning_rate": 9.733072916666665e-07,
+      "loss": 2.4437,
+      "mean_token_accuracy": 0.5663877908140421,
+      "num_tokens": 69821.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.1147512137889861,
+      "epoch": 0.1896633475580844,
+      "grad_norm": 3.241530656814575,
+      "learning_rate": 9.66796875e-07,
+      "loss": 2.3645,
+      "mean_token_accuracy": 0.5583278063684702,
+      "num_tokens": 77437.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.1382376968860626,
+      "epoch": 0.20862968231389284,
+      "grad_norm": 2.926927328109741,
+      "learning_rate": 9.602864583333332e-07,
+      "loss": 2.2085,
+      "mean_token_accuracy": 0.5798784755170345,
+      "num_tokens": 85618.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.190623751282692,
+      "epoch": 0.22759601706970128,
+      "grad_norm": 2.580003499984741,
+      "learning_rate": 9.537760416666665e-07,
+      "loss": 2.1356,
+      "mean_token_accuracy": 0.5675114955753088,
+      "num_tokens": 93792.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.2170004837214947,
+      "epoch": 0.24656235182550973,
+      "grad_norm": 2.2754576206207275,
+      "learning_rate": 9.472656249999999e-07,
+      "loss": 2.0575,
+      "mean_token_accuracy": 0.5808480866253376,
+      "num_tokens": 101636.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.2965586438775063,
+      "epoch": 0.2655286865813182,
+      "grad_norm": 2.0113515853881836,
+      "learning_rate": 9.407552083333333e-07,
+      "loss": 2.0583,
+      "mean_token_accuracy": 0.5819144621491432,
+      "num_tokens": 109232.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.3018298715353012,
+      "epoch": 0.2844950213371266,
+      "grad_norm": 1.8721307516098022,
+      "learning_rate": 9.342447916666665e-07,
+      "loss": 1.9126,
+      "mean_token_accuracy": 0.6024341452866793,
+      "num_tokens": 116924.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.3164579145610333,
+      "epoch": 0.30346135609293506,
+      "grad_norm": 1.9804903268814087,
+      "learning_rate": 9.27734375e-07,
+      "loss": 1.8649,
+      "mean_token_accuracy": 0.6070760123431682,
+      "num_tokens": 124256.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.321300919353962,
+      "epoch": 0.32242769084874345,
+      "grad_norm": 1.8772928714752197,
+      "learning_rate": 9.212239583333333e-07,
+      "loss": 1.8434,
+      "mean_token_accuracy": 0.6069471161812544,
+      "num_tokens": 131929.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.3300878189504146,
+      "epoch": 0.3413940256045519,
+      "grad_norm": 1.734480381011963,
+      "learning_rate": 9.147135416666666e-07,
+      "loss": 1.7704,
+      "mean_token_accuracy": 0.6136918149888515,
+      "num_tokens": 139796.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.2977866984903812,
+      "epoch": 0.36036036036036034,
+      "grad_norm": 1.4145355224609375,
+      "learning_rate": 9.082031249999999e-07,
+      "loss": 1.6739,
+      "mean_token_accuracy": 0.6313016630709172,
+      "num_tokens": 147946.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.3758595705032348,
+      "epoch": 0.3793266951161688,
+      "grad_norm": 1.5995664596557617,
+      "learning_rate": 9.016927083333333e-07,
+      "loss": 1.6861,
+      "mean_token_accuracy": 0.6132468611001969,
+      "num_tokens": 155548.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.3915398299694062,
+      "epoch": 0.39829302987197723,
+      "grad_norm": 1.450819969177246,
+      "learning_rate": 8.951822916666665e-07,
+      "loss": 1.7079,
+      "mean_token_accuracy": 0.6195833217352629,
+      "num_tokens": 163633.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.3885319493710995,
+      "epoch": 0.4172593646277857,
+      "grad_norm": 1.565523624420166,
+      "learning_rate": 8.88671875e-07,
+      "loss": 1.6055,
+      "mean_token_accuracy": 0.631564911082387,
+      "num_tokens": 171429.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.42000814974308,
+      "epoch": 0.4362256993835941,
+      "grad_norm": 1.415071725845337,
+      "learning_rate": 8.821614583333334e-07,
+      "loss": 1.5733,
+      "mean_token_accuracy": 0.6355191901326179,
+      "num_tokens": 178813.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.4173048675060271,
+      "epoch": 0.45519203413940257,
+      "grad_norm": 1.4940714836120605,
+      "learning_rate": 8.756510416666666e-07,
+      "loss": 1.5686,
+      "mean_token_accuracy": 0.6268294852226972,
+      "num_tokens": 186804.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.406810748577118,
+      "epoch": 0.474158368895211,
+      "grad_norm": 1.481840968132019,
+      "learning_rate": 8.69140625e-07,
+      "loss": 1.5083,
+      "mean_token_accuracy": 0.6454955622553825,
+      "num_tokens": 194400.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.4201976731419563,
+      "epoch": 0.49312470365101946,
+      "grad_norm": 1.303875207901001,
+      "learning_rate": 8.626302083333333e-07,
+      "loss": 1.5492,
+      "mean_token_accuracy": 0.6335972256958484,
+      "num_tokens": 202121.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.4369812428951263,
+      "epoch": 0.5120910384068279,
+      "grad_norm": 1.2758477926254272,
+      "learning_rate": 8.561197916666665e-07,
+      "loss": 1.4994,
+      "mean_token_accuracy": 0.6454178035259247,
+      "num_tokens": 209643.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.391933375597,
+      "epoch": 0.5310573731626363,
+      "grad_norm": 1.2780486345291138,
+      "learning_rate": 8.49609375e-07,
+      "loss": 1.4388,
+      "mean_token_accuracy": 0.6508051678538322,
+      "num_tokens": 217154.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.4164466500282287,
+      "epoch": 0.5500237079184448,
+      "grad_norm": 1.3106104135513306,
+      "learning_rate": 8.430989583333334e-07,
+      "loss": 1.4383,
+      "mean_token_accuracy": 0.6559075094759464,
+      "num_tokens": 224833.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.396108366549015,
+      "epoch": 0.5689900426742532,
+      "grad_norm": 1.1580867767333984,
+      "learning_rate": 8.365885416666666e-07,
+      "loss": 1.4295,
+      "mean_token_accuracy": 0.6560030743479729,
+      "num_tokens": 232528.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.3968449525535107,
+      "epoch": 0.5879563774300617,
+      "grad_norm": 1.4947290420532227,
+      "learning_rate": 8.30078125e-07,
+      "loss": 1.4474,
+      "mean_token_accuracy": 0.6532723367214203,
+      "num_tokens": 240340.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.3659315139055253,
+      "epoch": 0.6069227121858701,
+      "grad_norm": 1.2400017976760864,
+      "learning_rate": 8.235677083333333e-07,
+      "loss": 1.4045,
+      "mean_token_accuracy": 0.6663874570280314,
+      "num_tokens": 248482.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.3608342856168747,
+      "epoch": 0.6258890469416786,
+      "grad_norm": 1.1932809352874756,
+      "learning_rate": 8.170572916666666e-07,
+      "loss": 1.3297,
+      "mean_token_accuracy": 0.6729936949908734,
+      "num_tokens": 256146.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.369440621137619,
+      "epoch": 0.6448553816974869,
+      "grad_norm": 1.304652452468872,
+      "learning_rate": 8.10546875e-07,
+      "loss": 1.3665,
+      "mean_token_accuracy": 0.6693637847900391,
+      "num_tokens": 264115.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.3208621554076672,
+      "epoch": 0.6638217164532954,
+      "grad_norm": 1.253781795501709,
+      "learning_rate": 8.040364583333334e-07,
+      "loss": 1.32,
+      "mean_token_accuracy": 0.6763425137847662,
+      "num_tokens": 271660.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.322153516113758,
+      "epoch": 0.6827880512091038,
+      "grad_norm": 1.2030173540115356,
+      "learning_rate": 7.975260416666666e-07,
+      "loss": 1.3156,
+      "mean_token_accuracy": 0.6803543917834759,
+      "num_tokens": 279383.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.3421797141432763,
+      "epoch": 0.7017543859649122,
+      "grad_norm": 1.215598464012146,
+      "learning_rate": 7.91015625e-07,
+      "loss": 1.3249,
+      "mean_token_accuracy": 0.6731918178498745,
+      "num_tokens": 287094.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.3480857253074645,
+      "epoch": 0.7207207207207207,
+      "grad_norm": 1.3740663528442383,
+      "learning_rate": 7.845052083333333e-07,
+      "loss": 1.3468,
+      "mean_token_accuracy": 0.6654974155128002,
+      "num_tokens": 294860.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.3071250945329667,
+      "epoch": 0.7396870554765291,
+      "grad_norm": 1.37200129032135,
+      "learning_rate": 7.779947916666666e-07,
+      "loss": 1.3141,
+      "mean_token_accuracy": 0.6753718614578247,
+      "num_tokens": 302632.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.2966938570141793,
+      "epoch": 0.7586533902323376,
+      "grad_norm": 1.2524887323379517,
+      "learning_rate": 7.71484375e-07,
+      "loss": 1.2985,
+      "mean_token_accuracy": 0.6792728714644909,
+      "num_tokens": 310895.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.3287690579891205,
+      "epoch": 0.777619724988146,
+      "grad_norm": 1.4232484102249146,
+      "learning_rate": 7.649739583333334e-07,
+      "loss": 1.3058,
+      "mean_token_accuracy": 0.683712650835514,
+      "num_tokens": 318800.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.3336769036948681,
+      "epoch": 0.7965860597439545,
+      "grad_norm": 1.5561081171035767,
+      "learning_rate": 7.584635416666666e-07,
+      "loss": 1.3498,
+      "mean_token_accuracy": 0.6661995306611062,
+      "num_tokens": 326858.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.30598786175251,
+      "epoch": 0.8155523944997629,
+      "grad_norm": 1.274189829826355,
+      "learning_rate": 7.51953125e-07,
+      "loss": 1.2534,
+      "mean_token_accuracy": 0.6906878098845481,
+      "num_tokens": 334660.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.2970546074211597,
+      "epoch": 0.8345187292555714,
+      "grad_norm": 1.214945673942566,
+      "learning_rate": 7.454427083333333e-07,
+      "loss": 1.2538,
+      "mean_token_accuracy": 0.6886132195591926,
+      "num_tokens": 342557.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.3024328142404555,
+      "epoch": 0.8534850640113798,
+      "grad_norm": 1.4416435956954956,
+      "learning_rate": 7.389322916666666e-07,
+      "loss": 1.2292,
+      "mean_token_accuracy": 0.6991933032870292,
+      "num_tokens": 350090.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.2910141475498675,
+      "epoch": 0.8724513987671882,
+      "grad_norm": 1.3008499145507812,
+      "learning_rate": 7.32421875e-07,
+      "loss": 1.2491,
+      "mean_token_accuracy": 0.6929278194904327,
+      "num_tokens": 357993.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.280844808369875,
+      "epoch": 0.8914177335229967,
+      "grad_norm": 1.1629010438919067,
+      "learning_rate": 7.259114583333334e-07,
+      "loss": 1.2173,
+      "mean_token_accuracy": 0.7058381117880345,
+      "num_tokens": 365851.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.2642140217125415,
+      "epoch": 0.9103840682788051,
+      "grad_norm": 1.1701194047927856,
+      "learning_rate": 7.194010416666666e-07,
+      "loss": 1.1973,
+      "mean_token_accuracy": 0.7036448389291763,
+      "num_tokens": 373810.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.2932376384735107,
+      "epoch": 0.9293504030346136,
+      "grad_norm": 1.1186771392822266,
+      "learning_rate": 7.12890625e-07,
+      "loss": 1.2275,
+      "mean_token_accuracy": 0.6950574479997158,
+      "num_tokens": 381473.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.252262619137764,
+      "epoch": 0.948316737790422,
+      "grad_norm": 1.1925098896026611,
+      "learning_rate": 7.063802083333333e-07,
+      "loss": 1.1992,
+      "mean_token_accuracy": 0.7074964240193367,
+      "num_tokens": 389449.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.2449406683444977,
+      "epoch": 0.9672830725462305,
+      "grad_norm": 1.3617829084396362,
+      "learning_rate": 6.998697916666666e-07,
+      "loss": 1.1762,
+      "mean_token_accuracy": 0.7085213892161846,
+      "num_tokens": 397020.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.258817932009697,
+      "epoch": 0.9862494073020389,
+      "grad_norm": 1.1232348680496216,
+      "learning_rate": 6.93359375e-07,
+      "loss": 1.2082,
+      "mean_token_accuracy": 0.7091631673276424,
+      "num_tokens": 404954.0,
+      "step": 520
+    },
+    {
+      "entropy": 1.2346063224045005,
+      "epoch": 1.0037932669511618,
+      "grad_norm": 1.1821157932281494,
+      "learning_rate": 6.868489583333334e-07,
+      "loss": 1.1771,
+      "mean_token_accuracy": 0.7097870578636994,
+      "num_tokens": 411903.0,
+      "step": 530
+    },
+    {
+      "entropy": 1.233978260308504,
+      "epoch": 1.0227596017069702,
+      "grad_norm": 1.0421421527862549,
+      "learning_rate": 6.803385416666666e-07,
+      "loss": 1.2021,
+      "mean_token_accuracy": 0.7076010942459107,
+      "num_tokens": 419937.0,
+      "step": 540
+    },
+    {
+      "entropy": 1.197313880175352,
+      "epoch": 1.0417259364627787,
+      "grad_norm": 1.112593650817871,
+      "learning_rate": 6.73828125e-07,
+      "loss": 1.1401,
+      "mean_token_accuracy": 0.7142150066792965,
+      "num_tokens": 427521.0,
+      "step": 550
+    },
+    {
+      "entropy": 1.20088602155447,
+      "epoch": 1.0606922712185871,
+      "grad_norm": 1.0684127807617188,
+      "learning_rate": 6.673177083333333e-07,
+      "loss": 1.144,
+      "mean_token_accuracy": 0.7204408749938012,
+      "num_tokens": 435531.0,
+      "step": 560
+    },
+    {
+      "entropy": 1.2182291388511657,
+      "epoch": 1.0796586059743953,
+      "grad_norm": 1.0190104246139526,
+      "learning_rate": 6.608072916666666e-07,
+      "loss": 1.1465,
+      "mean_token_accuracy": 0.7094649709761143,
+      "num_tokens": 443298.0,
+      "step": 570
+    },
+    {
+      "entropy": 1.2136165104806422,
+      "epoch": 1.098624940730204,
+      "grad_norm": 1.413822889328003,
+      "learning_rate": 6.54296875e-07,
+      "loss": 1.1673,
+      "mean_token_accuracy": 0.7108944803476334,
+      "num_tokens": 451072.0,
+      "step": 580
+    },
+    {
+      "entropy": 1.2116065241396428,
+      "epoch": 1.1175912754860122,
+      "grad_norm": 1.0302143096923828,
+      "learning_rate": 6.477864583333334e-07,
+      "loss": 1.1624,
+      "mean_token_accuracy": 0.7103484153747559,
+      "num_tokens": 458701.0,
+      "step": 590
+    },
+    {
+      "entropy": 1.1983465410768985,
+      "epoch": 1.136557610241821,
+      "grad_norm": 1.1263726949691772,
+      "learning_rate": 6.412760416666666e-07,
+      "loss": 1.1779,
+      "mean_token_accuracy": 0.7140828415751457,
+      "num_tokens": 466421.0,
+      "step": 600
+    },
+    {
+      "entropy": 1.1892149433493615,
+      "epoch": 1.1555239449976291,
+      "grad_norm": 1.0706696510314941,
+      "learning_rate": 6.34765625e-07,
+      "loss": 1.1544,
+      "mean_token_accuracy": 0.7146961018443108,
+      "num_tokens": 474293.0,
+      "step": 610
+    },
+    {
+      "entropy": 1.1959151603281497,
+      "epoch": 1.1744902797534376,
+      "grad_norm": 1.028617024421692,
+      "learning_rate": 6.282552083333333e-07,
+      "loss": 1.1715,
+      "mean_token_accuracy": 0.7075264886021614,
+      "num_tokens": 482060.0,
+      "step": 620
+    },
+    {
+      "entropy": 1.186061894148588,
+      "epoch": 1.193456614509246,
+      "grad_norm": 0.8902014493942261,
+      "learning_rate": 6.217447916666666e-07,
+      "loss": 1.1364,
+      "mean_token_accuracy": 0.7179924137890339,
+      "num_tokens": 489638.0,
+      "step": 630
+    },
+    {
+      "entropy": 1.1988991744816304,
+      "epoch": 1.2124229492650545,
+      "grad_norm": 1.1088058948516846,
+      "learning_rate": 6.152343749999999e-07,
+      "loss": 1.1778,
+      "mean_token_accuracy": 0.705387695133686,
+      "num_tokens": 497269.0,
+      "step": 640
+    },
+    {
+      "entropy": 1.1572193197906018,
+      "epoch": 1.231389284020863,
+      "grad_norm": 1.0083446502685547,
+      "learning_rate": 6.087239583333334e-07,
+      "loss": 1.1343,
+      "mean_token_accuracy": 0.7159858278930187,
+      "num_tokens": 505048.0,
+      "step": 650
+    },
+    {
+      "entropy": 1.1919703222811222,
+      "epoch": 1.2503556187766713,
+      "grad_norm": 1.1133224964141846,
+      "learning_rate": 6.022135416666666e-07,
+      "loss": 1.1955,
+      "mean_token_accuracy": 0.7046414032578469,
+      "num_tokens": 513066.0,
+      "step": 660
+    },
+    {
+      "entropy": 1.1698683701455592,
+      "epoch": 1.2693219535324798,
+      "grad_norm": 1.1328527927398682,
+      "learning_rate": 5.95703125e-07,
+      "loss": 1.1461,
+      "mean_token_accuracy": 0.7104850746691227,
+      "num_tokens": 520866.0,
+      "step": 670
+    },
+    {
+      "entropy": 1.1433877781033517,
+      "epoch": 1.2882882882882882,
+      "grad_norm": 1.1092615127563477,
+      "learning_rate": 5.891927083333333e-07,
+      "loss": 1.0744,
+      "mean_token_accuracy": 0.7238348044455052,
+      "num_tokens": 528797.0,
+      "step": 680
+    },
+    {
+      "entropy": 1.1681585662066936,
+      "epoch": 1.3072546230440967,
+      "grad_norm": 0.9807539582252502,
+      "learning_rate": 5.826822916666666e-07,
+      "loss": 1.117,
+      "mean_token_accuracy": 0.7158304169774056,
+      "num_tokens": 536301.0,
+      "step": 690
+    },
+    {
+      "entropy": 1.1547326557338238,
+      "epoch": 1.3262209577999051,
+      "grad_norm": 1.101926565170288,
+      "learning_rate": 5.761718749999999e-07,
+      "loss": 1.1062,
+      "mean_token_accuracy": 0.7163656286895275,
+      "num_tokens": 543840.0,
+      "step": 700
+    },
+    {
+      "entropy": 1.1788968741893768,
+      "epoch": 1.3451872925557136,
+      "grad_norm": 1.1549991369247437,
+      "learning_rate": 5.696614583333334e-07,
+      "loss": 1.1142,
+      "mean_token_accuracy": 0.7148068331182003,
+      "num_tokens": 551411.0,
+      "step": 710
+    },
+    {
+      "entropy": 1.1192301869392396,
+      "epoch": 1.364153627311522,
+      "grad_norm": 1.110608696937561,
+      "learning_rate": 5.631510416666666e-07,
+      "loss": 1.1069,
+      "mean_token_accuracy": 0.7199234012514353,
+      "num_tokens": 559281.0,
+      "step": 720
+    },
+    {
+      "entropy": 1.150602648407221,
+      "epoch": 1.3831199620673305,
+      "grad_norm": 1.1103157997131348,
+      "learning_rate": 5.56640625e-07,
+      "loss": 1.1044,
+      "mean_token_accuracy": 0.7195970468223095,
+      "num_tokens": 566818.0,
+      "step": 730
+    },
+    {
+      "entropy": 1.1309661015868187,
+      "epoch": 1.402086296823139,
+      "grad_norm": 1.1398253440856934,
+      "learning_rate": 5.501302083333333e-07,
+      "loss": 1.0886,
+      "mean_token_accuracy": 0.722535713762045,
+      "num_tokens": 574633.0,
+      "step": 740
+    },
+    {
+      "entropy": 1.1047178946435452,
+      "epoch": 1.4210526315789473,
+      "grad_norm": 1.058017611503601,
+      "learning_rate": 5.436197916666666e-07,
+      "loss": 1.1284,
+      "mean_token_accuracy": 0.7201976045966149,
+      "num_tokens": 582679.0,
+      "step": 750
+    },
+    {
+      "entropy": 1.1124145589768886,
+      "epoch": 1.4400189663347558,
+      "grad_norm": 0.9302801489830017,
+      "learning_rate": 5.371093749999999e-07,
+      "loss": 1.1147,
+      "mean_token_accuracy": 0.7166914887726307,
+      "num_tokens": 590557.0,
+      "step": 760
+    },
+    {
+      "entropy": 1.1066633492708207,
+      "epoch": 1.4589853010905642,
+      "grad_norm": 1.0189460515975952,
+      "learning_rate": 5.305989583333334e-07,
+      "loss": 1.066,
+      "mean_token_accuracy": 0.7239819258451462,
+      "num_tokens": 598398.0,
+      "step": 770
+    },
+    {
+      "entropy": 1.1011511944234371,
+      "epoch": 1.4779516358463727,
+      "grad_norm": 1.0235430002212524,
+      "learning_rate": 5.240885416666666e-07,
+      "loss": 1.0808,
+      "mean_token_accuracy": 0.7260910376906395,
+      "num_tokens": 606364.0,
+      "step": 780
+    },
+    {
+      "entropy": 1.1028256572782993,
+      "epoch": 1.4969179706021811,
+      "grad_norm": 1.019422173500061,
+      "learning_rate": 5.17578125e-07,
+      "loss": 1.1052,
+      "mean_token_accuracy": 0.7217176735401154,
+      "num_tokens": 614105.0,
+      "step": 790
+    },
+    {
+      "entropy": 1.1365131378173827,
+      "epoch": 1.5158843053579896,
+      "grad_norm": 1.0515458583831787,
+      "learning_rate": 5.110677083333333e-07,
+      "loss": 1.1337,
+      "mean_token_accuracy": 0.7029635466635227,
+      "num_tokens": 621854.0,
+      "step": 800
+    },
+    {
+      "entropy": 1.1104629278182983,
+      "epoch": 1.534850640113798,
+      "grad_norm": 1.1318267583847046,
+      "learning_rate": 5.045572916666666e-07,
+      "loss": 1.0827,
+      "mean_token_accuracy": 0.7223107375204563,
+      "num_tokens": 629853.0,
+      "step": 810
+    },
+    {
+      "entropy": 1.130864578485489,
+      "epoch": 1.5538169748696065,
+      "grad_norm": 0.9779661893844604,
+      "learning_rate": 4.980468749999999e-07,
+      "loss": 1.1363,
+      "mean_token_accuracy": 0.7188829615712166,
+      "num_tokens": 637423.0,
+      "step": 820
+    },
+    {
+      "entropy": 1.0849515922367572,
+      "epoch": 1.572783309625415,
+      "grad_norm": 1.0097724199295044,
+      "learning_rate": 4.915364583333333e-07,
+      "loss": 1.0674,
+      "mean_token_accuracy": 0.7200270384550095,
+      "num_tokens": 645171.0,
+      "step": 830
+    },
+    {
+      "entropy": 1.0869166798889638,
+      "epoch": 1.5917496443812233,
+      "grad_norm": 0.9466545581817627,
+      "learning_rate": 4.850260416666666e-07,
+      "loss": 1.0693,
+      "mean_token_accuracy": 0.7244003631174565,
+      "num_tokens": 653064.0,
+      "step": 840
+    },
+    {
+      "entropy": 1.1223813638091087,
+      "epoch": 1.6107159791370318,
+      "grad_norm": 1.2643778324127197,
+      "learning_rate": 4.78515625e-07,
+      "loss": 1.1304,
+      "mean_token_accuracy": 0.7090435460209846,
+      "num_tokens": 661038.0,
+      "step": 850
+    },
+    {
+      "entropy": 1.1304323747754097,
+      "epoch": 1.6296823138928402,
+      "grad_norm": 1.177894115447998,
+      "learning_rate": 4.720052083333333e-07,
+      "loss": 1.1429,
+      "mean_token_accuracy": 0.7115202896296978,
+      "num_tokens": 669002.0,
+      "step": 860
+    },
+    {
+      "entropy": 1.1122850120067596,
+      "epoch": 1.6486486486486487,
+      "grad_norm": 1.061855673789978,
+      "learning_rate": 4.654947916666667e-07,
+      "loss": 1.0931,
+      "mean_token_accuracy": 0.7274209313094616,
+      "num_tokens": 676731.0,
+      "step": 870
+    },
+    {
+      "entropy": 1.1170439019799232,
+      "epoch": 1.6676149834044571,
+      "grad_norm": 1.0852574110031128,
+      "learning_rate": 4.58984375e-07,
+      "loss": 1.1216,
+      "mean_token_accuracy": 0.7111018508672714,
+      "num_tokens": 684255.0,
+      "step": 880
+    },
+    {
+      "entropy": 1.0897154599428176,
+      "epoch": 1.6865813181602656,
+      "grad_norm": 1.0128169059753418,
+      "learning_rate": 4.524739583333333e-07,
+      "loss": 1.067,
+      "mean_token_accuracy": 0.7212625838816166,
+      "num_tokens": 691876.0,
+      "step": 890
+    },
+    {
+      "entropy": 1.1157985217869282,
+      "epoch": 1.705547652916074,
+      "grad_norm": 1.164413332939148,
+      "learning_rate": 4.459635416666667e-07,
+      "loss": 1.111,
+      "mean_token_accuracy": 0.7152813680469989,
+      "num_tokens": 699586.0,
+      "step": 900
+    },
+    {
+      "entropy": 1.1163954354822636,
+      "epoch": 1.7245139876718825,
+      "grad_norm": 0.9794188737869263,
+      "learning_rate": 4.39453125e-07,
+      "loss": 1.1118,
+      "mean_token_accuracy": 0.7159987986087799,
+      "num_tokens": 707274.0,
+      "step": 910
+    },
+    {
+      "entropy": 1.1300695367157458,
+      "epoch": 1.7434803224276907,
+      "grad_norm": 0.9664438962936401,
+      "learning_rate": 4.329427083333333e-07,
+      "loss": 1.1452,
+      "mean_token_accuracy": 0.7118972539901733,
+      "num_tokens": 715626.0,
+      "step": 920
+    },
+    {
+      "entropy": 1.0764743611216545,
+      "epoch": 1.7624466571834994,
+      "grad_norm": 1.0159790515899658,
+      "learning_rate": 4.264322916666667e-07,
+      "loss": 1.0559,
+      "mean_token_accuracy": 0.7252737738192081,
+      "num_tokens": 723479.0,
+      "step": 930
+    },
+    {
+      "entropy": 1.1230794332921505,
+      "epoch": 1.7814129919393076,
+      "grad_norm": 1.0365698337554932,
+      "learning_rate": 4.19921875e-07,
+      "loss": 1.0958,
+      "mean_token_accuracy": 0.7190353393554687,
+      "num_tokens": 731226.0,
+      "step": 940
+    },
+    {
+      "entropy": 1.0928282856941223,
+      "epoch": 1.8003793266951162,
+      "grad_norm": 1.0494407415390015,
+      "learning_rate": 4.134114583333333e-07,
+      "loss": 1.0641,
+      "mean_token_accuracy": 0.7250751972198486,
+      "num_tokens": 739000.0,
+      "step": 950
+    },
+    {
+      "entropy": 1.0845353059470653,
+      "epoch": 1.8193456614509245,
+      "grad_norm": 1.3278709650039673,
+      "learning_rate": 4.069010416666667e-07,
+      "loss": 1.0595,
+      "mean_token_accuracy": 0.726919911801815,
+      "num_tokens": 746602.0,
+      "step": 960
+    },
+    {
+      "entropy": 1.0676946647465229,
+      "epoch": 1.8383119962067331,
+      "grad_norm": 1.053330421447754,
+      "learning_rate": 4.00390625e-07,
+      "loss": 1.0855,
+      "mean_token_accuracy": 0.7161190271377563,
+      "num_tokens": 754587.0,
+      "step": 970
+    },
+    {
+      "entropy": 1.0771495133638382,
+      "epoch": 1.8572783309625414,
+      "grad_norm": 1.1275739669799805,
+      "learning_rate": 3.938802083333333e-07,
+      "loss": 1.0597,
+      "mean_token_accuracy": 0.7268183678388596,
+      "num_tokens": 762321.0,
+      "step": 980
+    },
+    {
+      "entropy": 1.0400340281426907,
+      "epoch": 1.87624466571835,
+      "grad_norm": 1.0730770826339722,
+      "learning_rate": 3.8736979166666664e-07,
+      "loss": 1.0414,
+      "mean_token_accuracy": 0.7288486234843731,
+      "num_tokens": 770289.0,
+      "step": 990
+    },
+    {
+      "entropy": 1.0911719463765621,
+      "epoch": 1.8952110004741582,
+      "grad_norm": 1.155945062637329,
+      "learning_rate": 3.80859375e-07,
+      "loss": 1.0702,
+      "mean_token_accuracy": 0.7283292546868324,
+      "num_tokens": 778159.0,
+      "step": 1000
+    },
+    {
+      "entropy": 1.0983494065701962,
+      "epoch": 1.914177335229967,
+      "grad_norm": 0.9654527306556702,
+      "learning_rate": 3.743489583333333e-07,
+      "loss": 1.0865,
+      "mean_token_accuracy": 0.7174026511609555,
+      "num_tokens": 785898.0,
+      "step": 1010
+    },
+    {
+      "entropy": 1.0861414946615695,
+      "epoch": 1.9331436699857751,
+      "grad_norm": 1.131127119064331,
+      "learning_rate": 3.6783854166666664e-07,
+      "loss": 1.0835,
+      "mean_token_accuracy": 0.7165500566363334,
+      "num_tokens": 793613.0,
+      "step": 1020
+    },
+    {
+      "entropy": 1.0503841049969196,
+      "epoch": 1.9521100047415838,
+      "grad_norm": 1.0080134868621826,
+      "learning_rate": 3.61328125e-07,
+      "loss": 1.0242,
+      "mean_token_accuracy": 0.7282557636499405,
+      "num_tokens": 801437.0,
+      "step": 1030
+    },
+    {
+      "entropy": 1.0859679393470287,
+      "epoch": 1.971076339497392,
+      "grad_norm": 0.9900001287460327,
+      "learning_rate": 3.548177083333333e-07,
+      "loss": 1.1092,
+      "mean_token_accuracy": 0.7148670449852943,
+      "num_tokens": 809245.0,
+      "step": 1040
+    },
+    {
+      "entropy": 1.0815912760794162,
+      "epoch": 1.9900426742532007,
+      "grad_norm": 0.9635700583457947,
+      "learning_rate": 3.4830729166666664e-07,
+      "loss": 1.0859,
+      "mean_token_accuracy": 0.7190672904253006,
+      "num_tokens": 816867.0,
+      "step": 1050
+    },
+    {
+      "entropy": 1.0574141281682092,
+      "epoch": 2.0075865339023236,
+      "grad_norm": 1.097835898399353,
+      "learning_rate": 3.41796875e-07,
+      "loss": 1.0744,
+      "mean_token_accuracy": 0.7219740095976237,
+      "num_tokens": 823905.0,
+      "step": 1060
+    },
+    {
+      "entropy": 1.0721672594547271,
+      "epoch": 2.026552868658132,
+      "grad_norm": 0.985394299030304,
+      "learning_rate": 3.352864583333333e-07,
+      "loss": 1.0536,
+      "mean_token_accuracy": 0.7218626335263252,
+      "num_tokens": 831621.0,
+      "step": 1070
+    },
+    {
+      "entropy": 1.0695898160338402,
+      "epoch": 2.0455192034139404,
+      "grad_norm": 0.983224093914032,
+      "learning_rate": 3.2877604166666665e-07,
+      "loss": 1.0613,
+      "mean_token_accuracy": 0.725620323419571,
+      "num_tokens": 839561.0,
+      "step": 1080
+    },
+    {
+      "entropy": 1.054114455729723,
+      "epoch": 2.0644855381697487,
+      "grad_norm": 0.9491007328033447,
+      "learning_rate": 3.22265625e-07,
+      "loss": 1.0535,
+      "mean_token_accuracy": 0.7279413707554341,
+      "num_tokens": 847725.0,
+      "step": 1090
+    },
+    {
+      "entropy": 1.0666229009628296,
+      "epoch": 2.0834518729255573,
+      "grad_norm": 1.0753281116485596,
+      "learning_rate": 3.157552083333333e-07,
+      "loss": 1.0614,
+      "mean_token_accuracy": 0.7245796263217926,
+      "num_tokens": 855539.0,
+      "step": 1100
+    },
+    {
+      "entropy": 1.063340064138174,
+      "epoch": 2.1024182076813656,
+      "grad_norm": 1.0261895656585693,
+      "learning_rate": 3.0924479166666665e-07,
+      "loss": 1.0746,
+      "mean_token_accuracy": 0.7316980011761188,
+      "num_tokens": 863532.0,
+      "step": 1110
+    },
+    {
+      "entropy": 1.0661572083830833,
+      "epoch": 2.1213845424371742,
+      "grad_norm": 0.9890888333320618,
+      "learning_rate": 3.02734375e-07,
+      "loss": 1.0478,
+      "mean_token_accuracy": 0.7212581146508455,
+      "num_tokens": 871059.0,
+      "step": 1120
+    },
+    {
+      "entropy": 1.0444834858179093,
+      "epoch": 2.1403508771929824,
+      "grad_norm": 1.3268550634384155,
+      "learning_rate": 2.962239583333333e-07,
+      "loss": 1.0847,
+      "mean_token_accuracy": 0.7196186900138855,
+      "num_tokens": 878877.0,
+      "step": 1130
+    },
+    {
+      "entropy": 1.0576549746096133,
+      "epoch": 2.1593172119487907,
+      "grad_norm": 1.0555297136306763,
+      "learning_rate": 2.8971354166666665e-07,
+      "loss": 1.0529,
+      "mean_token_accuracy": 0.7290172964334488,
+      "num_tokens": 886515.0,
+      "step": 1140
+    },
+    {
+      "entropy": 1.0684065386652946,
+      "epoch": 2.1782835467045993,
+      "grad_norm": 1.1883318424224854,
+      "learning_rate": 2.83203125e-07,
+      "loss": 1.0726,
+      "mean_token_accuracy": 0.7240883596241474,
+      "num_tokens": 894300.0,
+      "step": 1150
+    },
+    {
+      "entropy": 1.057620432972908,
+      "epoch": 2.197249881460408,
+      "grad_norm": 1.0103634595870972,
+      "learning_rate": 2.766927083333333e-07,
+      "loss": 1.0704,
+      "mean_token_accuracy": 0.7233097150921821,
+      "num_tokens": 902075.0,
+      "step": 1160
+    },
+    {
+      "entropy": 1.0865561179816723,
+      "epoch": 2.2162162162162162,
+      "grad_norm": 1.123908519744873,
+      "learning_rate": 2.7018229166666665e-07,
+      "loss": 1.0793,
+      "mean_token_accuracy": 0.7224407941102982,
+      "num_tokens": 909708.0,
+      "step": 1170
+    },
+    {
+      "entropy": 1.075062484294176,
+      "epoch": 2.2351825509720245,
+      "grad_norm": 1.0090774297714233,
+      "learning_rate": 2.63671875e-07,
+      "loss": 1.0585,
+      "mean_token_accuracy": 0.7275661155581474,
+      "num_tokens": 917568.0,
+      "step": 1180
+    },
+    {
+      "entropy": 1.0581686094403266,
+      "epoch": 2.254148885727833,
+      "grad_norm": 1.0245497226715088,
+      "learning_rate": 2.571614583333333e-07,
+      "loss": 1.0439,
+      "mean_token_accuracy": 0.7295683771371841,
+      "num_tokens": 925359.0,
+      "step": 1190
+    },
+    {
+      "entropy": 1.0172349140048027,
+      "epoch": 2.273115220483642,
+      "grad_norm": 0.9528677463531494,
+      "learning_rate": 2.5065104166666665e-07,
+      "loss": 0.9964,
+      "mean_token_accuracy": 0.7368530400097371,
+      "num_tokens": 933085.0,
+      "step": 1200
+    },
+    {
+      "entropy": 1.0807716213166714,
+      "epoch": 2.29208155523945,
+      "grad_norm": 1.1366709470748901,
+      "learning_rate": 2.44140625e-07,
+      "loss": 1.0962,
+      "mean_token_accuracy": 0.7181029960513114,
+      "num_tokens": 941157.0,
+      "step": 1210
+    },
+    {
+      "entropy": 1.0512977451086045,
+      "epoch": 2.3110478899952582,
+      "grad_norm": 1.0394670963287354,
+      "learning_rate": 2.3763020833333333e-07,
+      "loss": 1.0311,
+      "mean_token_accuracy": 0.7289207331836224,
+      "num_tokens": 949006.0,
+      "step": 1220
+    },
+    {
+      "entropy": 1.0536490596830845,
+      "epoch": 2.330014224751067,
+      "grad_norm": 0.994644820690155,
+      "learning_rate": 2.3111979166666666e-07,
+      "loss": 1.0593,
+      "mean_token_accuracy": 0.7222072064876557,
+      "num_tokens": 956471.0,
+      "step": 1230
+    },
+    {
+      "entropy": 1.0701601654291153,
+      "epoch": 2.348980559506875,
+      "grad_norm": 1.160831332206726,
+      "learning_rate": 2.2460937499999998e-07,
+      "loss": 1.0706,
+      "mean_token_accuracy": 0.7194343417882919,
+      "num_tokens": 964391.0,
+      "step": 1240
+    },
+    {
+      "entropy": 1.059455419331789,
+      "epoch": 2.367946894262684,
+      "grad_norm": 0.9788382053375244,
+      "learning_rate": 2.1809895833333333e-07,
+      "loss": 1.0503,
+      "mean_token_accuracy": 0.7340390369296074,
+      "num_tokens": 972267.0,
+      "step": 1250
+    },
+    {
+      "entropy": 1.0449399091303349,
+      "epoch": 2.386913229018492,
+      "grad_norm": 1.0266343355178833,
+      "learning_rate": 2.1158854166666666e-07,
+      "loss": 1.0663,
+      "mean_token_accuracy": 0.7310878671705723,
+      "num_tokens": 980032.0,
+      "step": 1260
+    },
+    {
+      "entropy": 1.0491505213081838,
+      "epoch": 2.4058795637743007,
+      "grad_norm": 1.081602931022644,
+      "learning_rate": 2.0507812499999998e-07,
+      "loss": 1.0266,
+      "mean_token_accuracy": 0.726220253109932,
+      "num_tokens": 987459.0,
+      "step": 1270
+    },
+    {
+      "entropy": 1.04006467461586,
+      "epoch": 2.424845898530109,
+      "grad_norm": 1.0115092992782593,
+      "learning_rate": 1.9856770833333333e-07,
+      "loss": 1.0734,
+      "mean_token_accuracy": 0.7200278639793396,
+      "num_tokens": 995352.0,
+      "step": 1280
+    },
+    {
+      "entropy": 1.057293500751257,
+      "epoch": 2.4438122332859176,
+      "grad_norm": 1.0309503078460693,
+      "learning_rate": 1.9205729166666666e-07,
+      "loss": 1.0801,
+      "mean_token_accuracy": 0.7221748150885106,
+      "num_tokens": 1003216.0,
+      "step": 1290
+    },
+    {
+      "entropy": 1.0883753821253777,
+      "epoch": 2.462778568041726,
+      "grad_norm": 1.0229616165161133,
+      "learning_rate": 1.8554687499999998e-07,
+      "loss": 1.1076,
+      "mean_token_accuracy": 0.7148709982633591,
+      "num_tokens": 1011281.0,
+      "step": 1300
+    },
+    {
+      "entropy": 1.0345959886908531,
+      "epoch": 2.4817449027975345,
+      "grad_norm": 1.2275679111480713,
+      "learning_rate": 1.7903645833333334e-07,
+      "loss": 1.0265,
+      "mean_token_accuracy": 0.7251479253172874,
+      "num_tokens": 1019124.0,
+      "step": 1310
+    },
+    {
+      "entropy": 1.0503775343298911,
+      "epoch": 2.5007112375533427,
+      "grad_norm": 0.9931260943412781,
+      "learning_rate": 1.7252604166666666e-07,
+      "loss": 1.0592,
+      "mean_token_accuracy": 0.7259611785411835,
+      "num_tokens": 1027135.0,
+      "step": 1320
+    },
+    {
+      "entropy": 1.0528467200696467,
+      "epoch": 2.5196775723091513,
+      "grad_norm": 1.1964516639709473,
+      "learning_rate": 1.6601562499999999e-07,
+      "loss": 1.065,
+      "mean_token_accuracy": 0.7194369584321976,
+      "num_tokens": 1035098.0,
+      "step": 1330
+    },
+    {
+      "entropy": 1.0336830347776413,
+      "epoch": 2.5386439070649596,
+      "grad_norm": 1.0086219310760498,
+      "learning_rate": 1.5950520833333334e-07,
+      "loss": 1.0536,
+      "mean_token_accuracy": 0.7238833822309971,
+      "num_tokens": 1042919.0,
+      "step": 1340
+    },
+    {
+      "entropy": 1.0386064052581787,
+      "epoch": 2.5576102418207682,
+      "grad_norm": 1.0653084516525269,
+      "learning_rate": 1.5299479166666666e-07,
+      "loss": 1.027,
+      "mean_token_accuracy": 0.724561757594347,
+      "num_tokens": 1050649.0,
+      "step": 1350
+    },
+    {
+      "entropy": 1.065981125831604,
+      "epoch": 2.5765765765765765,
+      "grad_norm": 0.961837887763977,
+      "learning_rate": 1.46484375e-07,
+      "loss": 1.0915,
+      "mean_token_accuracy": 0.719189016520977,
+      "num_tokens": 1058569.0,
+      "step": 1360
+    },
+    {
+      "entropy": 1.0508228421211243,
+      "epoch": 2.595542911332385,
+      "grad_norm": 1.0830740928649902,
+      "learning_rate": 1.3997395833333334e-07,
+      "loss": 1.0482,
+      "mean_token_accuracy": 0.7348687440156937,
+      "num_tokens": 1066238.0,
+      "step": 1370
+    },
+    {
+      "entropy": 1.0544643633067607,
+      "epoch": 2.6145092460881934,
+      "grad_norm": 1.2174203395843506,
+      "learning_rate": 1.3346354166666666e-07,
+      "loss": 1.0435,
+      "mean_token_accuracy": 0.7309793442487716,
+      "num_tokens": 1074128.0,
+      "step": 1380
+    },
+    {
+      "entropy": 1.0388936534523965,
+      "epoch": 2.633475580844002,
+      "grad_norm": 1.0965008735656738,
+      "learning_rate": 1.26953125e-07,
+      "loss": 1.0296,
+      "mean_token_accuracy": 0.7309447333216668,
+      "num_tokens": 1082019.0,
+      "step": 1390
+    },
+    {
+      "entropy": 1.0290747590363025,
+      "epoch": 2.6524419155998102,
+      "grad_norm": 0.9872783422470093,
+      "learning_rate": 1.2044270833333331e-07,
+      "loss": 1.0335,
+      "mean_token_accuracy": 0.7279187783598899,
+      "num_tokens": 1089670.0,
+      "step": 1400
+    },
+    {
+      "entropy": 1.032333093881607,
+      "epoch": 2.671408250355619,
+      "grad_norm": 1.2452492713928223,
+      "learning_rate": 1.1393229166666667e-07,
+      "loss": 1.0232,
+      "mean_token_accuracy": 0.7304965436458588,
+      "num_tokens": 1097534.0,
+      "step": 1410
+    },
+    {
+      "entropy": 1.0229677349328994,
+      "epoch": 2.690374585111427,
+      "grad_norm": 1.234789490699768,
+      "learning_rate": 1.0742187499999999e-07,
+      "loss": 0.9894,
+      "mean_token_accuracy": 0.7343516901135445,
+      "num_tokens": 1104867.0,
+      "step": 1420
+    },
+    {
+      "entropy": 1.0563533768057822,
+      "epoch": 2.709340919867236,
+      "grad_norm": 0.9225372672080994,
+      "learning_rate": 1.0091145833333333e-07,
+      "loss": 1.0469,
+      "mean_token_accuracy": 0.7215077750384807,
+      "num_tokens": 1112409.0,
+      "step": 1430
+    },
+    {
+      "entropy": 1.0457329712808132,
+      "epoch": 2.728307254623044,
+      "grad_norm": 1.178075909614563,
+      "learning_rate": 9.440104166666667e-08,
+      "loss": 1.0233,
+      "mean_token_accuracy": 0.7347801722586155,
+      "num_tokens": 1119960.0,
+      "step": 1440
+    },
+    {
+      "entropy": 1.0247016452252864,
+      "epoch": 2.7472735893788527,
+      "grad_norm": 1.252854585647583,
+      "learning_rate": 8.789062499999999e-08,
+      "loss": 1.0095,
+      "mean_token_accuracy": 0.7328019939363003,
+      "num_tokens": 1127729.0,
+      "step": 1450
+    },
+    {
+      "entropy": 1.0175456829369067,
+      "epoch": 2.766239924134661,
+      "grad_norm": 1.0133498907089233,
+      "learning_rate": 8.138020833333333e-08,
+      "loss": 0.9892,
+      "mean_token_accuracy": 0.7394115671515464,
+      "num_tokens": 1135124.0,
+      "step": 1460
+    },
+    {
+      "entropy": 1.023030236363411,
+      "epoch": 2.7852062588904696,
+      "grad_norm": 1.1670140027999878,
+      "learning_rate": 7.486979166666667e-08,
+      "loss": 1.0009,
+      "mean_token_accuracy": 0.736342516541481,
+      "num_tokens": 1142969.0,
+      "step": 1470
+    },
+    {
+      "entropy": 1.0272307373583316,
+      "epoch": 2.804172593646278,
+      "grad_norm": 1.0572019815444946,
+      "learning_rate": 6.8359375e-08,
+      "loss": 1.0154,
+      "mean_token_accuracy": 0.7248648099601269,
+      "num_tokens": 1150592.0,
+      "step": 1480
+    },
+    {
+      "entropy": 1.0305047407746315,
+      "epoch": 2.823138928402086,
+      "grad_norm": 0.9308680295944214,
+      "learning_rate": 6.184895833333333e-08,
+      "loss": 1.0401,
+      "mean_token_accuracy": 0.7297416269779206,
+      "num_tokens": 1158674.0,
+      "step": 1490
+    },
+    {
+      "entropy": 1.069483907520771,
+      "epoch": 2.8421052631578947,
+      "grad_norm": 1.0559190511703491,
+      "learning_rate": 5.5338541666666664e-08,
+      "loss": 1.0788,
+      "mean_token_accuracy": 0.7130757793784142,
+      "num_tokens": 1166476.0,
+      "step": 1500
+    },
+    {
+      "entropy": 1.060176931321621,
+      "epoch": 2.8610715979137034,
+      "grad_norm": 1.2278039455413818,
+      "learning_rate": 4.8828124999999996e-08,
+      "loss": 1.1011,
+      "mean_token_accuracy": 0.7208118259906768,
+      "num_tokens": 1174300.0,
+      "step": 1510
+    },
+    {
+      "entropy": 1.0859525471925735,
+      "epoch": 2.8800379326695116,
+      "grad_norm": 1.0776472091674805,
+      "learning_rate": 4.2317708333333334e-08,
+      "loss": 1.1047,
+      "mean_token_accuracy": 0.715139028429985,
+      "num_tokens": 1182207.0,
+      "step": 1520
+    },
+    {
+      "entropy": 1.0568951837718488,
+      "epoch": 2.89900426742532,
+      "grad_norm": 1.1124858856201172,
+      "learning_rate": 3.5807291666666666e-08,
+      "loss": 1.0779,
+      "mean_token_accuracy": 0.7272238045930862,
+      "num_tokens": 1190118.0,
+      "step": 1530
+    },
+    {
+      "entropy": 1.0444737546145917,
+      "epoch": 2.9179706021811285,
+      "grad_norm": 1.2949162721633911,
+      "learning_rate": 2.9296874999999997e-08,
+      "loss": 1.0474,
+      "mean_token_accuracy": 0.7282775707542897,
+      "num_tokens": 1197864.0,
+      "step": 1540
+    },
+    {
+      "entropy": 1.0213447205722332,
+      "epoch": 2.936936936936937,
+      "grad_norm": 1.0395339727401733,
+      "learning_rate": 2.2786458333333332e-08,
+      "loss": 1.0082,
+      "mean_token_accuracy": 0.7363046444952488,
+      "num_tokens": 1205487.0,
+      "step": 1550
+    },
+    {
+      "entropy": 1.0271850168704986,
+      "epoch": 2.9559032716927454,
+      "grad_norm": 1.1394912004470825,
+      "learning_rate": 1.6276041666666667e-08,
+      "loss": 1.0238,
+      "mean_token_accuracy": 0.7308688588440418,
+      "num_tokens": 1213173.0,
+      "step": 1560
+    },
+    {
+      "entropy": 1.052078378200531,
+      "epoch": 2.9748696064485536,
+      "grad_norm": 1.111386775970459,
+      "learning_rate": 9.765624999999999e-09,
+      "loss": 1.0644,
+      "mean_token_accuracy": 0.7222773849964141,
+      "num_tokens": 1220993.0,
+      "step": 1570
+    },
+    {
+      "entropy": 1.070343980193138,
+      "epoch": 2.9938359412043622,
+      "grad_norm": 1.00790274143219,
+      "learning_rate": 3.2552083333333333e-09,
+      "loss": 1.0744,
+      "mean_token_accuracy": 0.7226169481873512,
+      "num_tokens": 1228824.0,
+      "step": 1580
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 1584,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4.999591237681152e+16,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

refusal-only-lora/checkpoint-1584/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36492051829b8ea0cb9bbd2945d49516e1af3ee03d355e12c6e414a818c5c5d5
+size 6417

refusal-only-lora/checkpoint-528/README.md

@@ -0,0 +1,209 @@
+---
+base_model: meta-llama/Llama-2-7b-chat-hf
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:meta-llama/Llama-2-7b-chat-hf
+- 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 0.18.1

refusal-only-lora/checkpoint-528/adapter_config.json

@@ -0,0 +1,46 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": null,
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-chat-hf",
+  "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": 128,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.18.1",
+  "qalora_group_size": 16,
+  "r": 64,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "v_proj",
+    "k_proj",
+    "o_proj",
+    "up_proj",
+    "gate_proj",
+    "q_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

refusal-only-lora/checkpoint-528/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8156962dd0cc1eb3cffb66decf7765a11d4ce72f2e6cb1f05ef3b3803a392cec
+size 639691872

refusal-only-lora/checkpoint-528/chat_template.jinja

@@ -0,0 +1 @@
+{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if loop.index0 == 0 and system_message != false %}{% set content = '<<SYS>>\n' + system_message + '\n<</SYS>>\n\n' + message['content'] %}{% else %}{% set content = message['content'] %}{% endif %}{% if message['role'] == 'user' %}{{ bos_token + '[INST] ' + content.strip() + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ ' '  + content.strip() + ' ' + eos_token }}{% endif %}{% endfor %}

refusal-only-lora/checkpoint-528/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8dd58a1cb9c40c8757268c6dfc3644c2bc9d0c7126a2d4a2f4fb95f425425bcd
+size 1279641507

refusal-only-lora/checkpoint-528/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab3db75b536cc459dc955e92af301958f46fb60600a2f9199c930a7b36d147a3
+size 14645

refusal-only-lora/checkpoint-528/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bcecc4b41f0572663eed2af60cc028a078509b376b79de535f8eb23b3c72393
+size 1465

refusal-only-lora/checkpoint-528/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

refusal-only-lora/checkpoint-528/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

refusal-only-lora/checkpoint-528/tokenizer_config.json

@@ -0,0 +1,43 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": null,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "extra_special_tokens": {},
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

refusal-only-lora/checkpoint-528/trainer_state.json

@@ -0,0 +1,554 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 528,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 0.8800771124660969,
+      "epoch": 0.01896633475580844,
+      "grad_norm": 3.348053455352783,
+      "learning_rate": 1.875e-07,
+      "loss": 2.892,
+      "mean_token_accuracy": 0.5558312267065049,
+      "num_tokens": 7568.0,
+      "step": 10
+    },
+    {
+      "entropy": 0.9028674259781837,
+      "epoch": 0.03793266951161688,
+      "grad_norm": 3.5438897609710693,
+      "learning_rate": 3.958333333333333e-07,
+      "loss": 2.84,
+      "mean_token_accuracy": 0.5506744537502527,
+      "num_tokens": 15418.0,
+      "step": 20
+    },
+    {
+      "entropy": 0.8906873933970928,
+      "epoch": 0.05689900426742532,
+      "grad_norm": 4.670811653137207,
+      "learning_rate": 6.041666666666666e-07,
+      "loss": 2.8846,
+      "mean_token_accuracy": 0.5502553284168243,
+      "num_tokens": 23186.0,
+      "step": 30
+    },
+    {
+      "entropy": 0.8964283980429173,
+      "epoch": 0.07586533902323377,
+      "grad_norm": 4.167201995849609,
+      "learning_rate": 8.125e-07,
+      "loss": 2.8202,
+      "mean_token_accuracy": 0.5535532858222723,
+      "num_tokens": 30766.0,
+      "step": 40
+    },
+    {
+      "entropy": 0.9246485389769077,
+      "epoch": 0.0948316737790422,
+      "grad_norm": 4.584895133972168,
+      "learning_rate": 9.993489583333333e-07,
+      "loss": 2.703,
+      "mean_token_accuracy": 0.5526723157614469,
+      "num_tokens": 38564.0,
+      "step": 50
+    },
+    {
+      "entropy": 0.9249856673181057,
+      "epoch": 0.11379800853485064,
+      "grad_norm": 4.198783874511719,
+      "learning_rate": 9.928385416666666e-07,
+      "loss": 2.6924,
+      "mean_token_accuracy": 0.5561300627887249,
+      "num_tokens": 46612.0,
+      "step": 60
+    },
+    {
+      "entropy": 0.9432196334004402,
+      "epoch": 0.1327643432906591,
+      "grad_norm": 4.151094436645508,
+      "learning_rate": 9.86328125e-07,
+      "loss": 2.5988,
+      "mean_token_accuracy": 0.5674021277576685,
+      "num_tokens": 54312.0,
+      "step": 70
+    },
+    {
+      "entropy": 1.0061342589557172,
+      "epoch": 0.15173067804646753,
+      "grad_norm": 3.3446481227874756,
+      "learning_rate": 9.798177083333333e-07,
+      "loss": 2.5035,
+      "mean_token_accuracy": 0.565283863991499,
+      "num_tokens": 62281.0,
+      "step": 80
+    },
+    {
+      "entropy": 1.0555521003901958,
+      "epoch": 0.17069701280227595,
+      "grad_norm": 4.015676498413086,
+      "learning_rate": 9.733072916666665e-07,
+      "loss": 2.4437,
+      "mean_token_accuracy": 0.5663877908140421,
+      "num_tokens": 69821.0,
+      "step": 90
+    },
+    {
+      "entropy": 1.1147512137889861,
+      "epoch": 0.1896633475580844,
+      "grad_norm": 3.241530656814575,
+      "learning_rate": 9.66796875e-07,
+      "loss": 2.3645,
+      "mean_token_accuracy": 0.5583278063684702,
+      "num_tokens": 77437.0,
+      "step": 100
+    },
+    {
+      "entropy": 1.1382376968860626,
+      "epoch": 0.20862968231389284,
+      "grad_norm": 2.926927328109741,
+      "learning_rate": 9.602864583333332e-07,
+      "loss": 2.2085,
+      "mean_token_accuracy": 0.5798784755170345,
+      "num_tokens": 85618.0,
+      "step": 110
+    },
+    {
+      "entropy": 1.190623751282692,
+      "epoch": 0.22759601706970128,
+      "grad_norm": 2.580003499984741,
+      "learning_rate": 9.537760416666665e-07,
+      "loss": 2.1356,
+      "mean_token_accuracy": 0.5675114955753088,
+      "num_tokens": 93792.0,
+      "step": 120
+    },
+    {
+      "entropy": 1.2170004837214947,
+      "epoch": 0.24656235182550973,
+      "grad_norm": 2.2754576206207275,
+      "learning_rate": 9.472656249999999e-07,
+      "loss": 2.0575,
+      "mean_token_accuracy": 0.5808480866253376,
+      "num_tokens": 101636.0,
+      "step": 130
+    },
+    {
+      "entropy": 1.2965586438775063,
+      "epoch": 0.2655286865813182,
+      "grad_norm": 2.0113515853881836,
+      "learning_rate": 9.407552083333333e-07,
+      "loss": 2.0583,
+      "mean_token_accuracy": 0.5819144621491432,
+      "num_tokens": 109232.0,
+      "step": 140
+    },
+    {
+      "entropy": 1.3018298715353012,
+      "epoch": 0.2844950213371266,
+      "grad_norm": 1.8721307516098022,
+      "learning_rate": 9.342447916666665e-07,
+      "loss": 1.9126,
+      "mean_token_accuracy": 0.6024341452866793,
+      "num_tokens": 116924.0,
+      "step": 150
+    },
+    {
+      "entropy": 1.3164579145610333,
+      "epoch": 0.30346135609293506,
+      "grad_norm": 1.9804903268814087,
+      "learning_rate": 9.27734375e-07,
+      "loss": 1.8649,
+      "mean_token_accuracy": 0.6070760123431682,
+      "num_tokens": 124256.0,
+      "step": 160
+    },
+    {
+      "entropy": 1.321300919353962,
+      "epoch": 0.32242769084874345,
+      "grad_norm": 1.8772928714752197,
+      "learning_rate": 9.212239583333333e-07,
+      "loss": 1.8434,
+      "mean_token_accuracy": 0.6069471161812544,
+      "num_tokens": 131929.0,
+      "step": 170
+    },
+    {
+      "entropy": 1.3300878189504146,
+      "epoch": 0.3413940256045519,
+      "grad_norm": 1.734480381011963,
+      "learning_rate": 9.147135416666666e-07,
+      "loss": 1.7704,
+      "mean_token_accuracy": 0.6136918149888515,
+      "num_tokens": 139796.0,
+      "step": 180
+    },
+    {
+      "entropy": 1.2977866984903812,
+      "epoch": 0.36036036036036034,
+      "grad_norm": 1.4145355224609375,
+      "learning_rate": 9.082031249999999e-07,
+      "loss": 1.6739,
+      "mean_token_accuracy": 0.6313016630709172,
+      "num_tokens": 147946.0,
+      "step": 190
+    },
+    {
+      "entropy": 1.3758595705032348,
+      "epoch": 0.3793266951161688,
+      "grad_norm": 1.5995664596557617,
+      "learning_rate": 9.016927083333333e-07,
+      "loss": 1.6861,
+      "mean_token_accuracy": 0.6132468611001969,
+      "num_tokens": 155548.0,
+      "step": 200
+    },
+    {
+      "entropy": 1.3915398299694062,
+      "epoch": 0.39829302987197723,
+      "grad_norm": 1.450819969177246,
+      "learning_rate": 8.951822916666665e-07,
+      "loss": 1.7079,
+      "mean_token_accuracy": 0.6195833217352629,
+      "num_tokens": 163633.0,
+      "step": 210
+    },
+    {
+      "entropy": 1.3885319493710995,
+      "epoch": 0.4172593646277857,
+      "grad_norm": 1.565523624420166,
+      "learning_rate": 8.88671875e-07,
+      "loss": 1.6055,
+      "mean_token_accuracy": 0.631564911082387,
+      "num_tokens": 171429.0,
+      "step": 220
+    },
+    {
+      "entropy": 1.42000814974308,
+      "epoch": 0.4362256993835941,
+      "grad_norm": 1.415071725845337,
+      "learning_rate": 8.821614583333334e-07,
+      "loss": 1.5733,
+      "mean_token_accuracy": 0.6355191901326179,
+      "num_tokens": 178813.0,
+      "step": 230
+    },
+    {
+      "entropy": 1.4173048675060271,
+      "epoch": 0.45519203413940257,
+      "grad_norm": 1.4940714836120605,
+      "learning_rate": 8.756510416666666e-07,
+      "loss": 1.5686,
+      "mean_token_accuracy": 0.6268294852226972,
+      "num_tokens": 186804.0,
+      "step": 240
+    },
+    {
+      "entropy": 1.406810748577118,
+      "epoch": 0.474158368895211,
+      "grad_norm": 1.481840968132019,
+      "learning_rate": 8.69140625e-07,
+      "loss": 1.5083,
+      "mean_token_accuracy": 0.6454955622553825,
+      "num_tokens": 194400.0,
+      "step": 250
+    },
+    {
+      "entropy": 1.4201976731419563,
+      "epoch": 0.49312470365101946,
+      "grad_norm": 1.303875207901001,
+      "learning_rate": 8.626302083333333e-07,
+      "loss": 1.5492,
+      "mean_token_accuracy": 0.6335972256958484,
+      "num_tokens": 202121.0,
+      "step": 260
+    },
+    {
+      "entropy": 1.4369812428951263,
+      "epoch": 0.5120910384068279,
+      "grad_norm": 1.2758477926254272,
+      "learning_rate": 8.561197916666665e-07,
+      "loss": 1.4994,
+      "mean_token_accuracy": 0.6454178035259247,
+      "num_tokens": 209643.0,
+      "step": 270
+    },
+    {
+      "entropy": 1.391933375597,
+      "epoch": 0.5310573731626363,
+      "grad_norm": 1.2780486345291138,
+      "learning_rate": 8.49609375e-07,
+      "loss": 1.4388,
+      "mean_token_accuracy": 0.6508051678538322,
+      "num_tokens": 217154.0,
+      "step": 280
+    },
+    {
+      "entropy": 1.4164466500282287,
+      "epoch": 0.5500237079184448,
+      "grad_norm": 1.3106104135513306,
+      "learning_rate": 8.430989583333334e-07,
+      "loss": 1.4383,
+      "mean_token_accuracy": 0.6559075094759464,
+      "num_tokens": 224833.0,
+      "step": 290
+    },
+    {
+      "entropy": 1.396108366549015,
+      "epoch": 0.5689900426742532,
+      "grad_norm": 1.1580867767333984,
+      "learning_rate": 8.365885416666666e-07,
+      "loss": 1.4295,
+      "mean_token_accuracy": 0.6560030743479729,
+      "num_tokens": 232528.0,
+      "step": 300
+    },
+    {
+      "entropy": 1.3968449525535107,
+      "epoch": 0.5879563774300617,
+      "grad_norm": 1.4947290420532227,
+      "learning_rate": 8.30078125e-07,
+      "loss": 1.4474,
+      "mean_token_accuracy": 0.6532723367214203,
+      "num_tokens": 240340.0,
+      "step": 310
+    },
+    {
+      "entropy": 1.3659315139055253,
+      "epoch": 0.6069227121858701,
+      "grad_norm": 1.2400017976760864,
+      "learning_rate": 8.235677083333333e-07,
+      "loss": 1.4045,
+      "mean_token_accuracy": 0.6663874570280314,
+      "num_tokens": 248482.0,
+      "step": 320
+    },
+    {
+      "entropy": 1.3608342856168747,
+      "epoch": 0.6258890469416786,
+      "grad_norm": 1.1932809352874756,
+      "learning_rate": 8.170572916666666e-07,
+      "loss": 1.3297,
+      "mean_token_accuracy": 0.6729936949908734,
+      "num_tokens": 256146.0,
+      "step": 330
+    },
+    {
+      "entropy": 1.369440621137619,
+      "epoch": 0.6448553816974869,
+      "grad_norm": 1.304652452468872,
+      "learning_rate": 8.10546875e-07,
+      "loss": 1.3665,
+      "mean_token_accuracy": 0.6693637847900391,
+      "num_tokens": 264115.0,
+      "step": 340
+    },
+    {
+      "entropy": 1.3208621554076672,
+      "epoch": 0.6638217164532954,
+      "grad_norm": 1.253781795501709,
+      "learning_rate": 8.040364583333334e-07,
+      "loss": 1.32,
+      "mean_token_accuracy": 0.6763425137847662,
+      "num_tokens": 271660.0,
+      "step": 350
+    },
+    {
+      "entropy": 1.322153516113758,
+      "epoch": 0.6827880512091038,
+      "grad_norm": 1.2030173540115356,
+      "learning_rate": 7.975260416666666e-07,
+      "loss": 1.3156,
+      "mean_token_accuracy": 0.6803543917834759,
+      "num_tokens": 279383.0,
+      "step": 360
+    },
+    {
+      "entropy": 1.3421797141432763,
+      "epoch": 0.7017543859649122,
+      "grad_norm": 1.215598464012146,
+      "learning_rate": 7.91015625e-07,
+      "loss": 1.3249,
+      "mean_token_accuracy": 0.6731918178498745,
+      "num_tokens": 287094.0,
+      "step": 370
+    },
+    {
+      "entropy": 1.3480857253074645,
+      "epoch": 0.7207207207207207,
+      "grad_norm": 1.3740663528442383,
+      "learning_rate": 7.845052083333333e-07,
+      "loss": 1.3468,
+      "mean_token_accuracy": 0.6654974155128002,
+      "num_tokens": 294860.0,
+      "step": 380
+    },
+    {
+      "entropy": 1.3071250945329667,
+      "epoch": 0.7396870554765291,
+      "grad_norm": 1.37200129032135,
+      "learning_rate": 7.779947916666666e-07,
+      "loss": 1.3141,
+      "mean_token_accuracy": 0.6753718614578247,
+      "num_tokens": 302632.0,
+      "step": 390
+    },
+    {
+      "entropy": 1.2966938570141793,
+      "epoch": 0.7586533902323376,
+      "grad_norm": 1.2524887323379517,
+      "learning_rate": 7.71484375e-07,
+      "loss": 1.2985,
+      "mean_token_accuracy": 0.6792728714644909,
+      "num_tokens": 310895.0,
+      "step": 400
+    },
+    {
+      "entropy": 1.3287690579891205,
+      "epoch": 0.777619724988146,
+      "grad_norm": 1.4232484102249146,
+      "learning_rate": 7.649739583333334e-07,
+      "loss": 1.3058,
+      "mean_token_accuracy": 0.683712650835514,
+      "num_tokens": 318800.0,
+      "step": 410
+    },
+    {
+      "entropy": 1.3336769036948681,
+      "epoch": 0.7965860597439545,
+      "grad_norm": 1.5561081171035767,
+      "learning_rate": 7.584635416666666e-07,
+      "loss": 1.3498,
+      "mean_token_accuracy": 0.6661995306611062,
+      "num_tokens": 326858.0,
+      "step": 420
+    },
+    {
+      "entropy": 1.30598786175251,
+      "epoch": 0.8155523944997629,
+      "grad_norm": 1.274189829826355,
+      "learning_rate": 7.51953125e-07,
+      "loss": 1.2534,
+      "mean_token_accuracy": 0.6906878098845481,
+      "num_tokens": 334660.0,
+      "step": 430
+    },
+    {
+      "entropy": 1.2970546074211597,
+      "epoch": 0.8345187292555714,
+      "grad_norm": 1.214945673942566,
+      "learning_rate": 7.454427083333333e-07,
+      "loss": 1.2538,
+      "mean_token_accuracy": 0.6886132195591926,
+      "num_tokens": 342557.0,
+      "step": 440
+    },
+    {
+      "entropy": 1.3024328142404555,
+      "epoch": 0.8534850640113798,
+      "grad_norm": 1.4416435956954956,
+      "learning_rate": 7.389322916666666e-07,
+      "loss": 1.2292,
+      "mean_token_accuracy": 0.6991933032870292,
+      "num_tokens": 350090.0,
+      "step": 450
+    },
+    {
+      "entropy": 1.2910141475498675,
+      "epoch": 0.8724513987671882,
+      "grad_norm": 1.3008499145507812,
+      "learning_rate": 7.32421875e-07,
+      "loss": 1.2491,
+      "mean_token_accuracy": 0.6929278194904327,
+      "num_tokens": 357993.0,
+      "step": 460
+    },
+    {
+      "entropy": 1.280844808369875,
+      "epoch": 0.8914177335229967,
+      "grad_norm": 1.1629010438919067,
+      "learning_rate": 7.259114583333334e-07,
+      "loss": 1.2173,
+      "mean_token_accuracy": 0.7058381117880345,
+      "num_tokens": 365851.0,
+      "step": 470
+    },
+    {
+      "entropy": 1.2642140217125415,
+      "epoch": 0.9103840682788051,
+      "grad_norm": 1.1701194047927856,
+      "learning_rate": 7.194010416666666e-07,
+      "loss": 1.1973,
+      "mean_token_accuracy": 0.7036448389291763,
+      "num_tokens": 373810.0,
+      "step": 480
+    },
+    {
+      "entropy": 1.2932376384735107,
+      "epoch": 0.9293504030346136,
+      "grad_norm": 1.1186771392822266,
+      "learning_rate": 7.12890625e-07,
+      "loss": 1.2275,
+      "mean_token_accuracy": 0.6950574479997158,
+      "num_tokens": 381473.0,
+      "step": 490
+    },
+    {
+      "entropy": 1.252262619137764,
+      "epoch": 0.948316737790422,
+      "grad_norm": 1.1925098896026611,
+      "learning_rate": 7.063802083333333e-07,
+      "loss": 1.1992,
+      "mean_token_accuracy": 0.7074964240193367,
+      "num_tokens": 389449.0,
+      "step": 500
+    },
+    {
+      "entropy": 1.2449406683444977,
+      "epoch": 0.9672830725462305,
+      "grad_norm": 1.3617829084396362,
+      "learning_rate": 6.998697916666666e-07,
+      "loss": 1.1762,
+      "mean_token_accuracy": 0.7085213892161846,
+      "num_tokens": 397020.0,
+      "step": 510
+    },
+    {
+      "entropy": 1.258817932009697,
+      "epoch": 0.9862494073020389,
+      "grad_norm": 1.1232348680496216,
+      "learning_rate": 6.93359375e-07,
+      "loss": 1.2082,
+      "mean_token_accuracy": 0.7091631673276424,
+      "num_tokens": 404954.0,
+      "step": 520
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 1584,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "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": 1.666530412560384e+16,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

refusal-only-lora/checkpoint-528/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36492051829b8ea0cb9bbd2945d49516e1af3ee03d355e12c6e414a818c5c5d5
+size 6417

refusal-only-lora/run_config.json

@@ -0,0 +1,13 @@
+{
+  "model_name_or_path": "meta-llama/Llama-2-7b-chat-hf",
+  "num_train_epochs": 3.0,
+  "learning_rate": 1e-06,
+  "full_finetune": false,
+  "use_qlora": false,
+  "lora_r": 64,
+  "dataset": "HarethahMo/extended-refusal",
+  "max_steps": null,
+  "response_aspect": "refusal_only",
+  "train_examples": 4218,
+  "assistant_only_loss": false
+}

refusal-only-lora/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

refusal-only-lora/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

refusal-only-lora/tokenizer_config.json

@@ -0,0 +1,43 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": null,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "extra_special_tokens": {},
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

refusal-only-lora/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36492051829b8ea0cb9bbd2945d49516e1af3ee03d355e12c6e414a818c5c5d5
+size 6417