metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:400
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
widget:
  - source_sentence: >-
      Which specific areas of law are mentioned as being unaffected by this
      Regulation?
    sentences:
      - (4)
      - >-
        (45)



        Practices that are prohibited by Union law, including data protection
        law, non-discrimination law, consumer protection law, and competition
        law, should not be affected by this Regulation.













        (46)
      - "Union harmonisation legislation in an optimal manner. AI systems identified as high-risk should be limited to those that have a\_significant harmful impact on the health, safety and fundamental rights of persons in the Union and such limitation should minimise any potential restriction to international trade."
  - source_sentence: How does AI contribute to environmentally beneficial outcomes?
    sentences:
      - "AI is a\_fast evolving family of technologies that contributes to a\_wide array of economic, environmental and societal benefits across the entire spectrum of industries and social activities. By improving prediction, optimising operations and resource allocation, and personalising digital solutions available for individuals and organisations, the use of AI can provide key competitive advantages to undertakings and support socially and environmentally beneficial outcomes, for example in healthcare, agriculture, food safety, education and training, media, sports, culture, infrastructure management, energy, transport and logistics, public services, security, justice, resource and energy efficiency, environmental monitoring, the conservation"
      - "To mitigate the risks from high-risk AI systems placed on the market or put into service and to ensure a\_high level of trustworthiness, certain mandatory requirements should apply to high-risk AI systems, taking into account the intended purpose and the context of use of the AI system and according to the risk-management system to be established by the provider. The measures adopted by the providers to comply with the mandatory requirements of this Regulation should take into account the generally acknowledged state of the art on AI, be proportionate and effective to meet the objectives of this Regulation. Based on the New Legislative Framework, as clarified in Commission notice ‘The “Blue Guide” on the implementation of EU product rules"
      - "Having regard to the proposal from the European Commission,\n\n\nAfter transmission of the draft legislative act to the national parliaments,\n\n\nHaving regard to the opinion of the European Economic and Social Committee\_(1),\n\n\nHaving regard to the opinion of the European Central Bank\_(2),\n\n\nHaving regard to the opinion of the Committee of the Regions\_(3),\n\n\nActing in accordance with the ordinary legislative procedure\_(4),\n\nWhereas:\n\n\n\n\n\n\n\n(1)"
  - source_sentence: >-
      What is the role of the Commission in providing guidance for the
      implementation of conditions for non-high-risk AI systems?
    sentences:
      - >-
        of suspects should not be ignored, in particular the difficulty in
        obtaining meaningful information on the functioning of those systems and
        the resulting difficulty in challenging their results in court, in
        particular by natural persons under investigation.
      - "of the conditions referred to above should draw up documentation of the assessment before that system is placed on the market or put into service and should provide that documentation to national competent authorities upon request. Such a\_provider should be obliged to register the AI system in the EU database established under this Regulation. With a\_view to providing further guidance for the practical implementation of the conditions under which the AI systems listed in an annex to this Regulation are, on an exceptional basis, non-high-risk, the Commission should, after consulting the Board, provide guidelines specifying that practical implementation, completed by a\_comprehensive list of practical examples of use cases of AI systems that"
      - "completed human activity that may be relevant for the purposes of the high-risk uses listed in an annex to this Regulation. Considering those characteristics, the AI system provides only an additional layer to a\_human activity with consequently lowered risk. That condition would, for example, apply to AI systems that are intended to improve the language used in previously drafted documents, for example in relation to professional tone, academic style of language or by aligning text to a\_certain brand messaging. The third condition should be that the AI system is intended to detect decision-making patterns or deviations from prior decision-making patterns. The risk would be lowered because the use of the AI system follows a\_previously"
  - source_sentence: >-
      How does the context surrounding the number 39 influence its
      interpretation?
    sentences:
      - (39)
      - "requested by the European Parliament\_(6)."
      - "under the UN Convention relating to the Status of Refugees done at Geneva on 28\_July 1951 as amended by the Protocol of 31\_January 1967. Nor should they be used to in any way infringe on the principle of non-refoulement, or to deny safe and effective legal avenues into the territory of the Union, including the right to international protection."
  - source_sentence: >-
      How does the number 63 relate to the overall theme or subject being
      discussed?
    sentences:
      - (60)
      - (63)
      - >-
        The deployment of AI systems in education is important to promote
        high-quality digital education and training and to allow all learners
        and teachers to acquire and share the necessary digital skills and
        competences, including media literacy, and critical thinking, to take an
        active part in the economy, society, and in democratic processes.
        However, AI systems used in education or vocational training, in
        particular for determining access or admission, for assigning persons to
        educational and vocational training institutions or programmes at all
        levels, for evaluating learning outcomes of persons, for assessing the
        appropriate level of education for an individual and materially
        influencing the level of education and training that individuals
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy@1
  - cosine_accuracy@3
  - cosine_accuracy@5
  - cosine_accuracy@10
  - cosine_precision@1
  - cosine_precision@3
  - cosine_precision@5
  - cosine_precision@10
  - cosine_recall@1
  - cosine_recall@3
  - cosine_recall@5
  - cosine_recall@10
  - cosine_ndcg@10
  - cosine_mrr@10
  - cosine_map@100
model-index:
  - name: SentenceTransformer
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: Unknown
          type: unknown
        metrics:
          - type: cosine_accuracy@1
            value: 0.9583333333333334
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 1
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 1
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 1
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.9583333333333334
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.3333333333333333
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.19999999999999998
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.09999999999999999
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.9583333333333334
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 1
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 1
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 1
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.9791666666666666
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.9722222222222223
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.9722222222222222
            name: Cosine Map@100

SentenceTransformer

This is a sentence-transformers model trained. It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

Model Type: Sentence Transformer
Maximum Sequence Length: 512 tokens
Output Dimensionality: 1024 dimensions
Similarity Function: Cosine Similarity

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("MikeCraBash/legal-ft-1")
# Run inference
sentences = [
    'How does the number 63 relate to the overall theme or subject being discussed?',
    '(63)',
    '(60)',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Information Retrieval

Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.9583
cosine_accuracy@3	1.0
cosine_accuracy@5	1.0
cosine_accuracy@10	1.0
cosine_precision@1	0.9583
cosine_precision@3	0.3333
cosine_precision@5	0.2
cosine_precision@10	0.1
cosine_recall@1	0.9583
cosine_recall@3	1.0
cosine_recall@5	1.0
cosine_recall@10	1.0
cosine_ndcg@10	0.9792
cosine_mrr@10	0.9722
cosine_map@100	0.9722

Training Details

Training Dataset

Unnamed Dataset

Size: 400 training samples
Columns: sentence_0 and sentence_1
Approximate statistics based on the first 400 samples:
sentence_0 sentence_1
type string string
details
min: 10 tokens
mean: 20.45 tokens
max: 35 tokens

min: 5 tokens
mean: 93.01 tokens
max: 186 tokens

	sentence_0	sentence_1
type	string	string
details	min: 10 tokens mean: 20.45 tokens max: 35 tokens	min: 5 tokens mean: 93.01 tokens max: 186 tokens

Samples:

sentence_0	sentence_1
`What types of risk analytics are permitted according to the context provided?`	solely on profiling them or on assessing their personality traits and characteristics should be prohibited. In any case, that prohibition does not refer to or touch upon risk analytics that are not based on the profiling of individuals or on the personality traits and characteristics of individuals, such as AI systems using risk analytics to assess the likelihood of financial fraud by undertakings on the basis of suspicious transactions or risk analytic tools to predict the likelihood of the localisation of narcotics or illicit goods by customs authorities, for example on the basis of known trafficking routes.
`Why is profiling individuals based on their personality traits prohibited?`	solely on profiling them or on assessing their personality traits and characteristics should be prohibited. In any case, that prohibition does not refer to or touch upon risk analytics that are not based on the profiling of individuals or on the personality traits and characteristics of individuals, such as AI systems using risk analytics to assess the likelihood of financial fraud by undertakings on the basis of suspicious transactions or risk analytic tools to predict the likelihood of the localisation of narcotics or illicit goods by customs authorities, for example on the basis of known trafficking routes.
`What criteria determine whether an AI system is classified as high-risk?`	`of AI systems that are high-risk and use cases that are not.`

Loss: MatryoshkaLoss with these parameters:

{
    "loss": "MultipleNegativesRankingLoss",
    "matryoshka_dims": [
        768,
        512,
        256,
        128,
        64
    ],
    "matryoshka_weights": [
        1,
        1,
        1,
        1,
        1
    ],
    "n_dims_per_step": -1
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 10
per_device_eval_batch_size: 10
num_train_epochs: 10
multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 10
per_device_eval_batch_size: 10
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 1
eval_accumulation_steps: None
torch_empty_cache_steps: None
learning_rate: 5e-05
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1
num_train_epochs: 10
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.0
warmup_steps: 0
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
save_safetensors: True
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
use_ipex: False
bf16: False
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: None
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
debug: []
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_prefetch_factor: None
past_index: -1
disable_tqdm: False
remove_unused_columns: True
label_names: None
load_best_model_at_end: False
ignore_data_skip: False
fsdp: []
fsdp_min_num_params: 0
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
fsdp_transformer_layer_cls_to_wrap: None
accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch
optim_args: None
adafactor: False
group_by_length: False
length_column_name: length
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
dataloader_pin_memory: True
dataloader_persistent_workers: False
skip_memory_metrics: True
use_legacy_prediction_loop: False
push_to_hub: False
resume_from_checkpoint: None
hub_model_id: None
hub_strategy: every_save
hub_private_repo: None
hub_always_push: False
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
include_for_metrics: []
eval_do_concat_batches: True
fp16_backend: auto
push_to_hub_model_id: None
push_to_hub_organization: None
mp_parameters:
auto_find_batch_size: False
full_determinism: False
torchdynamo: None
ray_scope: last
ddp_timeout: 1800
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
dispatch_batches: None
split_batches: None
include_tokens_per_second: False
include_num_input_tokens_seen: False
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: False
use_liger_kernel: False
eval_use_gather_object: False
average_tokens_across_devices: False
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	cosine_ndcg@10
1.0	40	0.9715
1.25	50	0.9792
2.0	80	0.9715
2.5	100	0.9715
3.0	120	0.9715
3.75	150	0.9715
4.0	160	0.9792
5.0	200	0.9792
6.0	240	0.9688
6.25	250	0.9792
7.0	280	0.9715
7.5	300	0.9792
8.0	320	0.9792
8.75	350	0.9792
9.0	360	0.9792
10.0	400	0.9792

Framework Versions

Python: 3.11.11
Sentence Transformers: 3.4.1
Transformers: 4.48.2
PyTorch: 2.5.1+cu124
Accelerate: 1.3.0
Datasets: 3.2.0
Tokenizers: 0.21.0

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}