README.md

metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:210
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
base_model: Snowflake/snowflake-arctic-embed-l
widget:
  - source_sentence: >-
      What does maintenance refer to in the context of providing for another
      person?  
    sentences:
      - >-
        -M-

        Maintenance:  The f urnishing by one  person to another the means of
        living, or fo od, clothing,
      - >-
        income and expenses to determine if the debtor may proceed under Chapter
        7.

        Chapter 7 trustee

        A person appointed in a Chapter 7 case to represent the interests of the
        bankruptcy estate

        and the creditors. The trustee's responsibilities include reviewing the
        debtor's petition and

        schedules, liquidating the property of the estate, and making
        distributions to creditors. The

        trustee may also bring actions against creditors or the debtor to
        recover property of the

        bankruptcy estate.

        Chapter 9
      - >-
        -19-

        Trial De Novo:  A new trial (See 22NYCRR 28.12).

        -U-

        Undertaking:  Deposit of a sum of money or filing of a bond in court, to
        secure some actual or

        potential obligation.

        -V-

        Vacate:  To set aside or undo a previous action or order.

        Venire:  Technically, a writ summoning persons to court to act as
        jurors; popularly used as meaning

        the body of names thus summoned.

        Venue:  (a) Geographical place where some legal matter occurs or may be
        determined. (b) The

        geographical area within which a court has jurisdiction. It relates only
        to a place or territory within

        which either party may require a case to be tried. A defect in venue may
        be waived by the parties.
  - source_sentence: 'What does the term "Pro Se" refer to in a legal context?  '
    sentences:
      - >-
        Process:  A l egal means, such as a s ummons, used to s ubject a de
        fendant i n a l awsuit to the

        personal jur isdiction o f the c ourt; broa dly, r efers to all writs
        iss ued i n the c ourse of a  le gal

        proceeding - what is served to obtain jurisdiction.

        Pro Se (aka Self-Represented):  Appearing on one’s own behalf without an
        attorney.

        Purge:  To atone for or correct an offense, to submit to a court's
        mandate (i.e., to purge oneself

        of contempt of court).

        -Q-

        None.

        -R-

        Recuse:  To disqualify oneself as a judge.

        Redact:  To edit, revise or block out written text.

        Referee:  A person to whom a claim  pending in a court is referred by
        the court to take testimony,
      - >-
        -10-

        Hearing:  A pr eliminary examination where testimony is given and e
        vidence presented for the

        purpose of determining an issue of fact and reaching a decision on the
        basis of that evidence.

        Hearsay:  Testimony of a witness who relates not what he/she knows
        personally, but what others

        have told the witness, or what the witness has heard said by others; may
        be admissible or

        inadmissible in court depending upon rules of evidence.

        Hung Jury:  A jury whose members cannot reconcile their differences of
        opinion and thus cannot

        reach a verdict.

        -I-

        Impaneling:  The process by which jurors are selected and sworn to their
        task.

        Impleader:  An addition of another party to an action by the defendant,
        a “third party” claim.
      - >-
        -12-

        Jurisdiction, Subject Matter:  Whether the court has authority over the
        thing or right claimed by

        one party against another.

        Jury:  A prescribed number of persons selected according to law and
        sworn to make findings of

        fact.

        Jury (Advisory):  A body of jurors impaneled to hear a case in which the
        parties have no right to

        a jury trial - the judge remains solely responsible for the findings and
        may accept or reject the

        jury's verdict.

        Jury Instructions:  Directions given by the judge to the jury, at the
        beginning and end of trial.

        -K-

        None.

        -L-

        Laches:  The failure to diligently assert a right, which results in a
        refusal to allow the right to be

        asserted later.

        Legal Age:  Eighteen (18) years of age. See CPLR Section 1206.
  - source_sentence: What is the purpose of a Chapter 11 bankruptcy filing?
    sentences:
      - >-
        condemnation, i.e., the legal process by which real estate of a private
        owner is taken for public use

        without the owner's consent, but upon the award and payment of just
        compensation.

        Enjoin:  To require a person, by writ of injunction from a court of
        equity, to perform or to refrain

        from  or cease doing some act.

        Entry:  The formal filing of an order of judgment with the County Clerk.

        Equitable Action (Equity Matter):  An action which may be brought for
        the purpose of restraining
      - >-
        A legal claim.

        Chambers

        The offices of a judge and his or her staff.

        Chapter 11

        A reorganization bankruptcy, usually involving a corporation or
        partnership. A Chapter 11

        debtor usually proposes a plan of reorganization to keep its business
        alive and pay creditors

        over time. Individuals or people in business can also seek relief in
        Chapter 11.

        Chapter 12

        The chapter of the Bankruptcy Code providing for adjustment of debts of
        a "family farmer"

        or "family fisherman," as the terms are defined in the Bankruptcy Code.

        Chapter 13

        The chapter of the Bankruptcy Code providing for the adjustment of debts
        of an individual

        with regular income, often referred to as a "wage-earner" plan. Chapter
        13 allows a debtor
      - >-
        Conviction

        A judgment of guilt against a criminal defendant.

        Counsel

        Legal advice; a term also used to refer to the lawyers in a case.

        Count

        An allegation in an indictment or information, charging a defendant with
        a crime. An

        indictment or information may contain allegations that the defendant
        committed more

        than one crime. Each allegation is referred to as a count.

        Court

        Government entity authorized to resolve legal disputes. Judges sometimes
        use "court" to

        refer to themselves in the third person, as in "the court has read the
        briefs."

        Court reporter

        A person who makes a word-for-word record of what is said in court,
        generally by using a

        stenographic machine, shorthand or audio recording, and then produces a
        transcript of the
  - source_sentence: >-
      What types of property may a debtor be able to exempt under the homestead
      exemption?  
    sentences:
      - >-
        -2-

        Affidavit of Service:  An affidavit intended to certify or prove that 
        service of a writ, notice, or other

        document has been made.

        Affirm:  An act of declaring something to be true under the penalty of
        perjury by a person who

        conscientiously declines to take an oath for religious or other
        pertinent reasons; also attorneys are

        permitted to affirm rather than swear under oath.

        Affirmation:  A solemn and formal declaration under penalties of perjury
        that a statement is true,

        without an oath.

        Affirmed:  Upheld, agreed with (e.g.,The Appellate Court affirmed the
        judgment of the City Court);

        also means a challenge to a court decision or order was rejected.
      - >-
        A formal request for the protection of the federal bankruptcy laws.
        (There is an official form

        for bankruptcy petitions.)

        Bankruptcy trustee

        A private individual or corporation appointed in all Chapter 7 and
        Chapter 13 cases to

        represent the interests of the bankruptcy estate and the debtor's
        creditors.

        Bench trial

        A trial without a jury, in which the judge serves as the fact-finder.

        Brief

        A written statement submitted in a trial or appellate proceeding that
        explains one side's

        legal and factual arguments.

        Burden of proof

        The duty to prove disputed facts. In civil cases, a plaintiff generally
        has the burden of

        proving his or her case. In criminal cases, the government has the
        burden of proving the

        defendant's guilt. (See standard of proof.)
      - >-
        residence (homestead exemption), or some or all "tools of the trade"
        used by the debtor to

        make a living (i.e., auto tools for an auto mechanic or dental tools for
        a dentist). The

        availability and amount of property the debtor may exempt depends on the
        state the debtor

        lives in.

        F

        Face sheet filing

        A bankruptcy case filed either without schedules or with incomplete
        schedules listing few

        creditors and debts. (Face sheet filings are often made for the purpose
        of delaying an
  - source_sentence: >-
      How does a fraudulent transfer relate to a debtor's intent in bankruptcy
      cases?
    sentences:
      - >-
        Glossary of Legal Terms

        Find definitions of legal terms to help understand the federal

        court system.

        A

        Acquittal

        A jury verdict that a criminal defendant is not guilty, or the finding
        of a judge that the

        evidence is insufficient to support a conviction.

        Active judge

        A judge in the full-time service of the court. Compare to senior judge.

        Administrative Office of the United States Courts (AO)

        Enter legal term to search for definition

        Search
      - >-
        A serious crime, usually punishable by at least one year in prison.

        File

        To place a paper in the official custody of the clerk of court to enter
        into the files or records

        of a case.

        Fraudulent transfer

        A transfer of a debtor's property made with intent to defraud or for
        which the debtor

        receives less than the transferred property's value.

        Fresh start

        The characterization of a debtor's status after bankruptcy, i.e., free
        of most debts. (Giving

        debtors a fresh start is one purpose of the Bankruptcy Code.)

        G

        Grand jury

        A body of 16-23 citizens who listen to evidence of criminal allegations,
        which is presented by

        the prosecutors, and determine whether there is probable cause to
        believe an individual
      - >-
        -3-

        Argument:  A reason given in proof or rebuttal to persuade a judge or
        jury.

        At Issue:  Whenever the parties to an action come to a point in the
        pleadings or argument which

        is affirmed on one side and denied on the other, the points  are said to
        be "at issue".

        Attachment:  The taking of property into legal custody by an enforcement
        officer (See specialty

        section: Recovery of Chattel).

        Attestation:  The act of witnessing an instrument in writing at the
        request of the party making the

        instrument and signing  it as a witness.

        Attorney of Record:  Attorney whose name appears in the court’s  records
        or files of a case.

        Award:  A decision of an Arbitrator, judge or jury.

        -B-
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 based on Snowflake/snowflake-arctic-embed-l
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: Unknown
          type: unknown
        metrics:
          - type: cosine_accuracy@1
            value: 0.9318181818181818
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.9318181818181818
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.9545454545454546
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 1
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.9318181818181818
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.3106060606060606
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1909090909090909
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.09999999999999996
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.9318181818181818
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.9318181818181818
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.9545454545454546
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 1
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.9565434941101226
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.9438131313131314
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.9438131313131314
            name: Cosine Map@100

SentenceTransformer based on Snowflake/snowflake-arctic-embed-l

This is a sentence-transformers model finetuned from Snowflake/snowflake-arctic-embed-l. 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
• Base model: Snowflake/snowflake-arctic-embed-l
• 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("vin00d/snowflake-arctic-legal-ft-1")
# Run inference
sentences = [
    "How does a fraudulent transfer relate to a debtor's intent in bankruptcy cases?",
    "A serious crime, usually punishable by at least one year in prison.\nFile\nTo place a paper in the official custody of the clerk of court to enter into the files or records\nof a case.\nFraudulent transfer\nA transfer of a debtor's property made with intent to defraud or for which the debtor\nreceives less than the transferred property's value.\nFresh start\nThe characterization of a debtor's status after bankruptcy, i.e., free of most debts. (Giving\ndebtors a fresh start is one purpose of the Bankruptcy Code.)\nG\nGrand jury\nA body of 16-23 citizens who listen to evidence of criminal allegations, which is presented by\nthe prosecutors, and determine whether there is probable cause to believe an individual",
    '-3-\nArgument:  A reason given in proof or rebuttal to persuade a judge or jury.\nAt Issue:  Whenever the parties to an action come to a point in the pleadings or argument which\nis affirmed on one side and denied on the other, the points  are said to be "at issue".\nAttachment:  The taking of property into legal custody by an enforcement officer (See specialty\nsection: Recovery of Chattel).\nAttestation:  The act of witnessing an instrument in writing at the request of the party making the\ninstrument and signing  it as a witness.\nAttorney of Record:  Attorney whose name appears in the court’s  records or files of a case.\nAward:  A decision of an Arbitrator, judge or jury.\n-B-',
]
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	0.9318
cosine_accuracy@3	0.9318
cosine_accuracy@5	0.9545
cosine_accuracy@10	1.0
cosine_precision@1	0.9318
cosine_precision@3	0.3106
cosine_precision@5	0.1909
cosine_precision@10	0.1
cosine_recall@1	0.9318
cosine_recall@3	0.9318
cosine_recall@5	0.9545
cosine_recall@10	1.0
cosine_ndcg@10	0.9565
cosine_mrr@10	0.9438
cosine_map@100	0.9438

Training Details

Training Dataset

Unnamed Dataset

• Size: 210 training samples
• Columns: sentence_0 and sentence_1
• Approximate statistics based on the first 210 samples:

  	sentence_0	sentence_1
  type	string	string
  details	min: 9 tokens mean: 17.36 tokens max: 33 tokens	min: 4 tokens mean: 122.9 tokens max: 192 tokens

• Samples:

  sentence_0	sentence_1
  What is the purpose of the glossary of common legal terms provided in the context?	GLOSSARY ‐ COMMON LEGAL TERMS NOTE:  The following definitions are not legal definitions.  Rather, these definitions are intended to give you a general idea of the meanings of common legal words.  For  comprehensive Definitions of legal terms, you may wish to consult a legal dictionary  “Black’s Law Dictionary” is one such legal dictionary which is usually available at  most law libraries. This glossary of common legal terms is also available on‐line at: http://www.nycourts.gov/lawlibraries/glossary.shtml ADDITIONAL ON‐LINE RESOURCES: http://www.nolo.com/glossary.cfm  Nolo’s on‐line legal dictionary. http://www.law‐dictionary.org/ Free on‐line legal dictionary search engine. http://www.law.cornell.edu/wex
  Where can one find a comprehensive legal dictionary for more detailed definitions of legal terms?	GLOSSARY ‐ COMMON LEGAL TERMS NOTE:  The following definitions are not legal definitions.  Rather, these definitions are intended to give you a general idea of the meanings of common legal words.  For  comprehensive Definitions of legal terms, you may wish to consult a legal dictionary  “Black’s Law Dictionary” is one such legal dictionary which is usually available at  most law libraries. This glossary of common legal terms is also available on‐line at: http://www.nycourts.gov/lawlibraries/glossary.shtml ADDITIONAL ON‐LINE RESOURCES: http://www.nolo.com/glossary.cfm  Nolo’s on‐line legal dictionary. http://www.law‐dictionary.org/ Free on‐line legal dictionary search engine. http://www.law.cornell.edu/wex
  What organization maintains the legal dictionary and encyclopedia mentioned in the context?	Legal dictionary and encyclopedia maintained by the Legal Information Institute at Cornell Law School.

• 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	21	0.9240
2.0	42	0.9628
2.3810	50	0.9628
3.0	63	0.9502
4.0	84	0.9569
4.7619	100	0.9563
5.0	105	0.9556
6.0	126	0.9569
7.0	147	0.9555
7.1429	150	0.9555
8.0	168	0.9565
9.0	189	0.9565
9.5238	200	0.9565
10.0	210	0.9565

Framework Versions

• Python: 3.11.11
• Sentence Transformers: 3.4.1
• Transformers: 4.48.3
• PyTorch: 2.5.1+cu124
• Accelerate: 1.3.0
• Datasets: 3.3.2
• 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}
}