IoannisKat1 commited on
Commit
46d92b5
·
verified ·
1 Parent(s): 42557c6

Add finetuned model

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +3 -0
  2. README.md +0 -0
  3. checkpoint-14/1_Pooling/config.json +10 -0
  4. checkpoint-14/README.md +1637 -0
  5. checkpoint-14/config.json +27 -0
  6. checkpoint-14/config_sentence_transformers.json +14 -0
  7. checkpoint-14/model.safetensors +3 -0
  8. checkpoint-14/modules.json +20 -0
  9. checkpoint-14/optimizer.pt +3 -0
  10. checkpoint-14/rng_state.pth +3 -0
  11. checkpoint-14/scheduler.pt +3 -0
  12. checkpoint-14/sentence_bert_config.json +4 -0
  13. checkpoint-14/sentencepiece.bpe.model +3 -0
  14. checkpoint-14/special_tokens_map.json +51 -0
  15. checkpoint-14/tokenizer.json +3 -0
  16. checkpoint-14/tokenizer_config.json +55 -0
  17. checkpoint-14/trainer_state.json +337 -0
  18. checkpoint-14/training_args.bin +3 -0
  19. checkpoint-28/1_Pooling/config.json +10 -0
  20. checkpoint-28/README.md +1651 -0
  21. checkpoint-28/config.json +27 -0
  22. checkpoint-28/config_sentence_transformers.json +14 -0
  23. checkpoint-28/model.safetensors +3 -0
  24. checkpoint-28/modules.json +20 -0
  25. checkpoint-28/optimizer.pt +3 -0
  26. checkpoint-28/rng_state.pth +3 -0
  27. checkpoint-28/scheduler.pt +3 -0
  28. checkpoint-28/sentence_bert_config.json +4 -0
  29. checkpoint-28/sentencepiece.bpe.model +3 -0
  30. checkpoint-28/special_tokens_map.json +51 -0
  31. checkpoint-28/tokenizer.json +3 -0
  32. checkpoint-28/tokenizer_config.json +55 -0
  33. checkpoint-28/trainer_state.json +631 -0
  34. checkpoint-28/training_args.bin +3 -0
  35. checkpoint-35/1_Pooling/config.json +10 -0
  36. checkpoint-35/README.md +1658 -0
  37. checkpoint-35/config.json +27 -0
  38. checkpoint-35/config_sentence_transformers.json +14 -0
  39. checkpoint-35/model.safetensors +3 -0
  40. checkpoint-35/modules.json +20 -0
  41. checkpoint-35/optimizer.pt +3 -0
  42. checkpoint-35/rng_state.pth +3 -0
  43. checkpoint-35/scheduler.pt +3 -0
  44. checkpoint-35/sentence_bert_config.json +4 -0
  45. checkpoint-35/sentencepiece.bpe.model +3 -0
  46. checkpoint-35/special_tokens_map.json +51 -0
  47. checkpoint-35/tokenizer.json +3 -0
  48. checkpoint-35/tokenizer_config.json +55 -0
  49. checkpoint-35/trainer_state.json +778 -0
  50. checkpoint-35/training_args.bin +3 -0
.gitattributes CHANGED
@@ -40,3 +40,6 @@ tokenizer.json filter=lfs diff=lfs merge=lfs -text
40
  checkpoint-39/tokenizer.json filter=lfs diff=lfs merge=lfs -text
41
  checkpoint-52/tokenizer.json filter=lfs diff=lfs merge=lfs -text
42
  checkpoint-65/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 
 
 
 
40
  checkpoint-39/tokenizer.json filter=lfs diff=lfs merge=lfs -text
41
  checkpoint-52/tokenizer.json filter=lfs diff=lfs merge=lfs -text
42
  checkpoint-65/tokenizer.json filter=lfs diff=lfs merge=lfs -text
43
+ checkpoint-14/tokenizer.json filter=lfs diff=lfs merge=lfs -text
44
+ checkpoint-28/tokenizer.json filter=lfs diff=lfs merge=lfs -text
45
+ checkpoint-35/tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
The diff for this file is too large to render. See raw diff
 
checkpoint-14/1_Pooling/config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "word_embedding_dimension": 1024,
3
+ "pooling_mode_cls_token": false,
4
+ "pooling_mode_mean_tokens": true,
5
+ "pooling_mode_max_tokens": false,
6
+ "pooling_mode_mean_sqrt_len_tokens": false,
7
+ "pooling_mode_weightedmean_tokens": false,
8
+ "pooling_mode_lasttoken": false,
9
+ "include_prompt": true
10
+ }
checkpoint-14/README.md ADDED
@@ -0,0 +1,1637 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license: apache-2.0
5
+ tags:
6
+ - sentence-transformers
7
+ - sentence-similarity
8
+ - feature-extraction
9
+ - dense
10
+ - generated_from_trainer
11
+ - dataset_size:82
12
+ - loss:MatryoshkaLoss
13
+ - loss:MultipleNegativesRankingLoss
14
+ base_model: intfloat/multilingual-e5-large
15
+ widget:
16
+ - source_sentence: When did the victims give away credentials?
17
+ sentences:
18
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
19
+
20
+
21
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
22
+ benefit, causes damage to another’s property by persuading someone to act, omit,
23
+ or tolerate something through the knowing misrepresentation of false facts as
24
+ true, or through the unlawful concealment or suppression of true facts, shall
25
+ be punished by imprisonment of at least three months, and if the damage caused
26
+ is particularly large, by imprisonment of at least two years."
27
+
28
+
29
+ From this provision it follows that, for the crime of fraud to be established,
30
+ the following elements are required:
31
+
32
+
33
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
34
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
35
+
36
+
37
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
38
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
39
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
40
+ to themselves or another; and
41
+
42
+
43
+ c) Damage to another person’s property, as defined under civil law, which must
44
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
45
+ not required that the person deceived and the person who suffered the damage be
46
+ the same individual.
47
+
48
+
49
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
50
+ relating to the past or present, and not to those that will occur in the future,
51
+ such as mere promises or contractual obligations. However, when such promises
52
+ or obligations are accompanied by false assurances and representations of other
53
+ false facts referring to the present or the past, in such a manner as to create
54
+ the impression of future fulfillment based on a false present situation fabricated
55
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
56
+ the crime of fraud is established.
57
+
58
+
59
+ The term “property” refers to the totality of a person’s economic assets that
60
+ possess monetary value, while damage to property means its reduction—specifically,
61
+ the difference between the monetary value the property had before the disposition
62
+ caused by the fraudulent conduct and the value remaining after it. Property damage
63
+ exists even if the victim possesses an active claim for restitution.
64
+
65
+
66
+ The time of commission of the fraud is considered to be the moment when the perpetrator
67
+ acted and completed their fraudulent conduct, namely when they made the false
68
+ representations that deceived the victim or a third party. Any subsequent moment
69
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
70
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
71
+ - 'Voice phishing involves manipulating victims over the phone. Attackers pose as
72
+ bank officials or authorities and use intimidation to extract financial details.
73
+
74
+
75
+ Scenario:
76
+
77
+ - Victims are coerced into giving away PINs, passwords, or other credentials under
78
+ false pretenses of legal or financial emergencies.'
79
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
80
+
81
+
82
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
83
+ benefit, causes damage to another’s property by persuading someone to act, omit,
84
+ or tolerate something through the knowing misrepresentation of false facts as
85
+ true, or through the unlawful concealment or suppression of true facts, shall
86
+ be punished by imprisonment of at least three months, and if the damage caused
87
+ is particularly large, by imprisonment of at least two years."
88
+
89
+
90
+ From this provision, it follows that, for the crime of fraud to be established,
91
+ the following elements are required:
92
+
93
+
94
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
95
+ pecuniary benefit, without requiring that the benefit actually materialize;
96
+
97
+
98
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
99
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
100
+ is deceived and performs an act, omission, or acquiescence; and
101
+
102
+
103
+ c) Damage to another’s property, according to civil law, which must be causally
104
+ connected to the perpetrator’s deceptive acts or omissions. It is not required
105
+ that the deceived person and the person who suffered the loss be the same.
106
+
107
+
108
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
109
+ relating to the past or present, and not to those that will occur in the future,
110
+ such as mere promises or contractual obligations. However, when such promises
111
+ or obligations are accompanied by false assurances and representations of other
112
+ false facts relating to the present or the past, in such a way as to create the
113
+ impression of future fulfillment, based on a false present situation fabricated
114
+ by the perpetrator—who has already made the decision not to fulfill their obligation—then
115
+ the crime of fraud is established.
116
+
117
+
118
+ The term “property” denotes the totality of a person’s economic assets possessing
119
+ monetary value, while damage to property refers to its reduction—specifically,
120
+ the difference between the property’s monetary value before the disposition caused
121
+ by the fraudulent conduct and its value afterward. Property damage exists even
122
+ if the victim has an active claim for its restitution.
123
+
124
+
125
+ The time of commission of fraud is considered to be the moment when the perpetrator
126
+ acted and completed the deceptive conduct, that is, when they made the false representations
127
+ which deceived the victim or a third party. Any later time at which the victim’s
128
+ financial loss occurred—thus completing the fraud—or the time when the harmful
129
+ act or omission of the deceived person took place, is irrelevant.
130
+
131
+
132
+ The reference to multiple modes of commission of fraud (i.e., both the misrepresentation
133
+ of false facts and the concealment of true ones) may create ambiguity and contradiction,
134
+ unless it is made clear from the overall findings that the offense was committed
135
+ in one particular manner, and that the reference to the other merely serves to
136
+ define the intent (mens rea) of the perpetrator—specifically, that the representations
137
+ were false.
138
+
139
+
140
+ Furthermore, a conviction must contain the specific and well-reasoned justification
141
+ required by Articles 93 paragraph 3 of the Constitution and 139 of the Code of
142
+ Criminal Procedure. The absence of such reasoning constitutes grounds for cassation
143
+ (appeal) under Article 510 paragraph 1(d) of the Code of Criminal Procedure, when
144
+ the judgment does not set out, with clarity, completeness, and consistency, the
145
+ factual circumstances established by the evidence, upon which the court based
146
+ its findings regarding the objective and subjective elements of the offense, the
147
+ evidence supporting those findings, and the legal reasoning through which those
148
+ facts were subsumed under the applicable substantive criminal provision.
149
+
150
+
151
+ For the existence of such reasoning, the explanatory and operative parts of the
152
+ decision may complement each other, as they form a single, unified whole.
153
+
154
+
155
+ The existence of intent (dolus) does not generally need to be specially justified,
156
+ since it is inherent in the will to bring about the factual circumstances constituting
157
+ the objective elements of the offense, and it is presumed from their realization
158
+ in each particular case—unless the law requires additional elements for criminal
159
+ liability, such as the act being committed with knowledge of a specific circumstance
160
+ (direct intent) or with the pursuit of a further purpose, i.e., the achievement
161
+ of an additional result (offenses requiring a special subjective element).
162
+
163
+
164
+ Furthermore, under Article 510 paragraph 1(e) of the Code of Criminal Procedure,
165
+ a misapplication of substantive criminal law also constitutes grounds for cassation.
166
+ Such misapplication occurs when the trial court incorrectly applies the law to
167
+ the facts it has found to be true, or when the violation occurs indirectly, namely
168
+ when the reasoning of the judgment—comprising the combination of its factual and
169
+ operative parts and relating to the elements and identity of the offense—contains
170
+ ambiguities, contradictions, or logical gaps, rendering it impossible to verify,
171
+ on appeal, whether the law was applied correctly. In such cases, the judgment
172
+ lacks a lawful basis.'
173
+ - source_sentence: What must be the outcome of the deception in relation to property
174
+ damage?
175
+ sentences:
176
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
177
+
178
+
179
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
180
+ benefit, causes damage to another’s property by persuading someone to act, omit,
181
+ or tolerate something through the knowing misrepresentation of false facts as
182
+ true, or through the unlawful concealment or suppression of true facts, shall
183
+ be punished by imprisonment of at least three months, and if the damage caused
184
+ is particularly large, by imprisonment of at least two years."
185
+
186
+
187
+ From this provision, it follows that, for the crime of fraud to be established,
188
+ the following elements are required:
189
+
190
+
191
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
192
+ pecuniary benefit, without requiring that the benefit actually materialize;
193
+
194
+
195
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
196
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
197
+ is deceived and performs an act, omission, or acquiescence; and
198
+
199
+
200
+ c) Damage to another’s property, according to civil law, which must be causally
201
+ connected to the perpetrator’s deceptive acts or omissions. It is not required
202
+ that the deceived person and the person who suffered the loss be the same.
203
+
204
+
205
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
206
+ relating to the past or present, and not to those that will occur in the future,
207
+ such as mere promises or contractual obligations. However, when such promises
208
+ or obligations are accompanied by false assurances and representations of other
209
+ false facts relating to the present or the past, in such a way as to create the
210
+ impression of future fulfillment, based on a false present situation fabricated
211
+ by the perpetrator—who has already made the decision not to fulfill their obligation—then
212
+ the crime of fraud is established.
213
+
214
+
215
+ The term “property” denotes the totality of a person’s economic assets possessing
216
+ monetary value, while damage to property refers to its reduction—specifically,
217
+ the difference between the property’s monetary value before the disposition caused
218
+ by the fraudulent conduct and its value afterward. Property damage exists even
219
+ if the victim has an active claim for its restitution.
220
+
221
+
222
+ The time of commission of fraud is considered to be the moment when the perpetrator
223
+ acted and completed the deceptive conduct, that is, when they made the false representations
224
+ which deceived the victim or a third party. Any later time at which the victim’s
225
+ financial loss occurred—thus completing the fraud—or the time when the harmful
226
+ act or omission of the deceived person took place, is irrelevant.
227
+
228
+
229
+ The reference to multiple modes of commission of fraud (i.e., both the misrepresentation
230
+ of false facts and the concealment of true ones) may create ambiguity and contradiction,
231
+ unless it is made clear from the overall findings that the offense was committed
232
+ in one particular manner, and that the reference to the other merely serves to
233
+ define the intent (mens rea) of the perpetrator—specifically, that the representations
234
+ were false.
235
+
236
+
237
+ Furthermore, a conviction must contain the specific and well-reasoned justification
238
+ required by Articles 93 paragraph 3 of the Constitution and 139 of the Code of
239
+ Criminal Procedure. The absence of such reasoning constitutes grounds for cassation
240
+ (appeal) under Article 510 paragraph 1(d) of the Code of Criminal Procedure, when
241
+ the judgment does not set out, with clarity, completeness, and consistency, the
242
+ factual circumstances established by the evidence, upon which the court based
243
+ its findings regarding the objective and subjective elements of the offense, the
244
+ evidence supporting those findings, and the legal reasoning through which those
245
+ facts were subsumed under the applicable substantive criminal provision.
246
+
247
+
248
+ For the existence of such reasoning, the explanatory and operative parts of the
249
+ decision may complement each other, as they form a single, unified whole.
250
+
251
+
252
+ The existence of intent (dolus) does not generally need to be specially justified,
253
+ since it is inherent in the will to bring about the factual circumstances constituting
254
+ the objective elements of the offense, and it is presumed from their realization
255
+ in each particular case—unless the law requires additional elements for criminal
256
+ liability, such as the act being committed with knowledge of a specific circumstance
257
+ (direct intent) or with the pursuit of a further purpose, i.e., the achievement
258
+ of an additional result (offenses requiring a special subjective element).
259
+
260
+
261
+ Furthermore, under Article 510 paragraph 1(e) of the Code of Criminal Procedure,
262
+ a misapplication of substantive criminal law also constitutes grounds for cassation.
263
+ Such misapplication occurs when the trial court incorrectly applies the law to
264
+ the facts it has found to be true, or when the violation occurs indirectly, namely
265
+ when the reasoning of the judgment—comprising the combination of its factual and
266
+ operative parts and relating to the elements and identity of the offense—contains
267
+ ambiguities, contradictions, or logical gaps, rendering it impossible to verify,
268
+ on appeal, whether the law was applied correctly. In such cases, the judgment
269
+ lacks a lawful basis.'
270
+ - 'According to Article 386 paragraph 1 of the Greek Penal Code,
271
+
272
+
273
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
274
+ benefit, causes damage to another’s property by persuading someone to act, omit,
275
+ or tolerate something through the knowing misrepresentation of false facts as
276
+ true, or through the unlawful concealment or suppression of true facts, shall
277
+ be punished by imprisonment of at least three months, and if the damage caused
278
+ is particularly large, by imprisonment of at least two years."
279
+
280
+
281
+ From these provisions, it follows that, for the crime of fraud to be established,
282
+ the following elements are required:
283
+
284
+
285
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
286
+ pecuniary benefit;
287
+
288
+
289
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
290
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
291
+ is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves
292
+ or another; and
293
+
294
+
295
+ c) Damage to another’s property, as defined under civil law, which must be causally
296
+ connected to the perpetrator’s deceptive acts.
297
+
298
+
299
+ From the above provisions, it is deduced that the crime of fraud is established
300
+ both objectively and subjectively through the knowing misrepresentation of false
301
+ facts as true, or the unlawful concealment or suppression of true ones, by which
302
+ another person is deceived and, as a result, performs an act, omission, or acquiescence
303
+ involving a disposition of property that directly and necessarily causes financial
304
+ damage to the deceived person or another, with the intent that the perpetrator
305
+ or another gain an unlawful benefit. It is irrelevant whether this intended benefit
306
+ was ultimately achieved.
307
+
308
+
309
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
310
+ relating to the past or present, and not to those expected to occur in the future,
311
+ such as mere promises or contractual obligations. The false fact must have existed
312
+ in the past or must be a present circumstance at the time it is asserted, and
313
+ cannot relate to the future.
314
+
315
+
316
+ However, when future circumstances—that is, promises or contractual obligations—are
317
+ accompanied by false assurances and representations of other false facts referring
318
+ to the present or past, in such a way as to create the impression of future fulfillment,
319
+ based on a false present situation or supposed ability of the perpetrator, who
320
+ had already made the decision not to fulfill their obligation, then the crime
321
+ of fraud is established.'
322
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
323
+
324
+
325
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
326
+ benefit, causes damage to another person’s property by persuading someone to act,
327
+ omit, or tolerate something through the knowing misrepresentation of false facts
328
+ as true, or through the unlawful concealment or suppression of true facts, shall
329
+ be punished by imprisonment of at least three months, and if the damage caused
330
+ is particularly large, by imprisonment of at least two years."
331
+
332
+
333
+ From this provision, it follows that for the crime of fraud to be established,
334
+ the following elements are required:
335
+
336
+
337
+ a) Intent of the perpetrator to obtain for themselves or another an unlawful pecuniary
338
+ benefit, regardless of whether this benefit was actually realized;
339
+
340
+
341
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
342
+ or suppression of true facts, as a result of which, as a causal factor, someone
343
+ is deceived and acts in a way that is detrimental to themselves or another (by
344
+ an act, omission, or acquiescence); and
345
+
346
+
347
+ c) Damage to another’s property, in the sense recognized by civil law, which must
348
+ be causally linked to the fraudulent conduct (the deceptive act or omission of
349
+ the perpetrator) and to the resulting deception of the person who made the property
350
+ disposition. It is not required that the person deceived be the same person who
351
+ suffered the damage.
352
+
353
+
354
+ Property damage exists when there is a reduction or deterioration in the victim’s
355
+ assets, even if the victim has an active claim to restitution. However, as an
356
+ element of the objective aspect of the crime of fraud, the damage must be the
357
+ direct, necessary, and exclusive result of the property disposition—namely, the
358
+ act, omission, or acquiescence performed by the person deceived by the perpetrator’s
359
+ fraudulent conduct.
360
+
361
+
362
+ There must therefore be a causal connection between the perpetrator’s deceptive
363
+ behavior and the deception it caused, as well as between this deception and the
364
+ resulting property damage, which must be the direct, necessary, and exclusive
365
+ outcome of the deception and of the act, omission, or acquiescence of the deceived
366
+ person.
367
+
368
+
369
+ The term “facts” refers to real circumstances relating to the past or present,
370
+ and not to those expected to occur in the future, such as mere promises or contractual
371
+ obligations. However, when such promises or obligations are accompanied by false
372
+ assurances and representations of other false facts relating to the present or
373
+ the past, in such a way as to create the impression of future fulfillment, based
374
+ on the false present situation presented by a perpetrator who has already made
375
+ the decision not to fulfill their obligation, then the crime of fraud is established.
376
+
377
+
378
+ The time of commission of the fraud is considered to be the moment when the perpetrator
379
+ acted and completed their deceptive conduct—that is, when they made the false
380
+ representations that deceived the victim or a third party. Any later time at which
381
+ the victim’s financial loss actually occurred—thus completing the fraud—or the
382
+ time when the deceived person performed the harmful act or omission, is irrelevant.'
383
+ - source_sentence: How are victims tricked in email phishing scams?
384
+ sentences:
385
+ - 'According to Article 386 paragraph 1 of the Greek Penal Code,
386
+
387
+
388
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
389
+ benefit, causes damage to another’s property by persuading someone to act, omit,
390
+ or tolerate something through the knowing misrepresentation of false facts as
391
+ true, or through the unlawful concealment or suppression of true facts, shall
392
+ be punished by imprisonment of at least three months, and if the damage caused
393
+ is particularly large, by imprisonment of at least two years."
394
+
395
+
396
+ From these provisions, it follows that, for the crime of fraud to be established,
397
+ the following elements are required:
398
+
399
+
400
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
401
+ pecuniary benefit;
402
+
403
+
404
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
405
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
406
+ is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves
407
+ or another; and
408
+
409
+
410
+ c) Damage to another’s property, as defined under civil law, which must be causally
411
+ connected to the perpetrator’s deceptive acts.
412
+
413
+
414
+ From the above provisions, it is deduced that the crime of fraud is established
415
+ both objectively and subjectively through the knowing misrepresentation of false
416
+ facts as true, or the unlawful concealment or suppression of true ones, by which
417
+ another person is deceived and, as a result, performs an act, omission, or acquiescence
418
+ involving a disposition of property that directly and necessarily causes financial
419
+ damage to the deceived person or another, with the intent that the perpetrator
420
+ or another gain an unlawful benefit. It is irrelevant whether this intended benefit
421
+ was ultimately achieved.
422
+
423
+
424
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
425
+ relating to the past or present, and not to those expected to occur in the future,
426
+ such as mere promises or contractual obligations. The false fact must have existed
427
+ in the past or must be a present circumstance at the time it is asserted, and
428
+ cannot relate to the future.
429
+
430
+
431
+ However, when future circumstances—that is, promises or contractual obligations—are
432
+ accompanied by false assurances and representations of other false facts referring
433
+ to the present or past, in such a way as to create the impression of future fulfillment,
434
+ based on a false present situation or supposed ability of the perpetrator, who
435
+ had already made the decision not to fulfill their obligation, then the crime
436
+ of fraud is established.'
437
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
438
+
439
+
440
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
441
+ benefit, causes damage to another’s property by persuading someone to act, omit,
442
+ or tolerate something through the knowing misrepresentation of false facts as
443
+ true, or through the unlawful concealment or suppression of true facts, shall
444
+ be punished by imprisonment of at least three months, and if the damage caused
445
+ is particularly large, by imprisonment of at least two years."
446
+
447
+
448
+ From this provision it follows that, for the crime of fraud to be established,
449
+ the following elements are required:
450
+
451
+
452
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
453
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
454
+
455
+
456
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
457
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
458
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
459
+ to themselves or another; and
460
+
461
+
462
+ c) Damage to another person’s property, as defined under civil law, which must
463
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
464
+ not required that the person deceived and the person who suffered the damage be
465
+ the same individual.
466
+
467
+
468
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
469
+ relating to the past or present, and not to those that will occur in the future,
470
+ such as mere promises or contractual obligations. However, when such promises
471
+ or obligations are accompanied by false assurances and representations of other
472
+ false facts referring to the present or the past, in such a manner as to create
473
+ the impression of future fulfillment based on a false present situation fabricated
474
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
475
+ the crime of fraud is established.
476
+
477
+
478
+ The term “property” refers to the totality of a person’s economic assets that
479
+ possess monetary value, while damage to property means its reduction—specifically,
480
+ the difference between the monetary value the property had before the disposition
481
+ caused by the fraudulent conduct and the value remaining after it. Property damage
482
+ exists even if the victim possesses an active claim for restitution.
483
+
484
+
485
+ The time of commission of the fraud is considered to be the moment when the perpetrator
486
+ acted and completed their fraudulent conduct, namely when they made the false
487
+ representations that deceived the victim or a third party. Any subsequent moment
488
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
489
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
490
+ - 'Email phishing is a type of identity theft scam conducted via email or SMS. The
491
+ attacker uses social engineering tactics such as impersonating trusted entities
492
+ and inducing urgency. Victims are tricked into disclosing personal information
493
+ or downloading malware.
494
+
495
+
496
+ Scenarios:
497
+
498
+ - Scenario 1: Emails impersonating high-ranking executives accuse victims of crimes
499
+ to coerce them into revealing information or opening malware-laden attachments.
500
+
501
+ - Scenario 2: Emails/SMS from fake banks or authorities alert victims of data
502
+ breaches, directing them to spoofed websites to input credentials.
503
+
504
+ - Scenario 3: SMS messages deliver disguised malware apps that harvest sensitive
505
+ data.
506
+
507
+ - Scenario 4: SMS links lead to pharming sites that mimic trusted brands and steal
508
+ login data through fake pop-ups.'
509
+ - source_sentence: What circumstances do the term 'facts' refer to within the meaning
510
+ of the provision?
511
+ sentences:
512
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
513
+
514
+
515
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
516
+ benefit, causes damage to another person’s property by persuading someone to act,
517
+ omit, or tolerate something through the knowing misrepresentation of false facts
518
+ as true, or through the unlawful concealment or suppression of true facts, shall
519
+ be punished by imprisonment of at least three months, and if the damage caused
520
+ is particularly large, by imprisonment of at least two years."
521
+
522
+
523
+ From this provision, it follows that for the crime of fraud to be established,
524
+ the following elements are required:
525
+
526
+
527
+ a) Intent of the perpetrator to obtain for themselves or another an unlawful pecuniary
528
+ benefit, regardless of whether this benefit was actually realized;
529
+
530
+
531
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
532
+ or suppression of true facts, as a result of which, as a causal factor, someone
533
+ is deceived and acts in a way that is detrimental to themselves or another (by
534
+ an act, omission, or acquiescence); and
535
+
536
+
537
+ c) Damage to another’s property, in the sense recognized by civil law, which must
538
+ be causally linked to the fraudulent conduct (the deceptive act or omission of
539
+ the perpetrator) and to the resulting deception of the person who made the property
540
+ disposition. It is not required that the person deceived be the same person who
541
+ suffered the damage.
542
+
543
+
544
+ Property damage exists when there is a reduction or deterioration in the victim’s
545
+ assets, even if the victim has an active claim to restitution. However, as an
546
+ element of the objective aspect of the crime of fraud, the damage must be the
547
+ direct, necessary, and exclusive result of the property disposition—namely, the
548
+ act, omission, or acquiescence performed by the person deceived by the perpetrator’s
549
+ fraudulent conduct.
550
+
551
+
552
+ There must therefore be a causal connection between the perpetrator’s deceptive
553
+ behavior and the deception it caused, as well as between this deception and the
554
+ resulting property damage, which must be the direct, necessary, and exclusive
555
+ outcome of the deception and of the act, omission, or acquiescence of the deceived
556
+ person.
557
+
558
+
559
+ The term “facts” refers to real circumstances relating to the past or present,
560
+ and not to those expected to occur in the future, such as mere promises or contractual
561
+ obligations. However, when such promises or obligations are accompanied by false
562
+ assurances and representations of other false facts relating to the present or
563
+ the past, in such a way as to create the impression of future fulfillment, based
564
+ on the false present situation presented by a perpetrator who has already made
565
+ the decision not to fulfill their obligation, then the crime of fraud is established.
566
+
567
+
568
+ The time of commission of the fraud is considered to be the moment when the perpetrator
569
+ acted and completed their deceptive conduct—that is, when they made the false
570
+ representations that deceived the victim or a third party. Any later time at which
571
+ the victim’s financial loss actually occurred—thus completing the fraud—or the
572
+ time when the deceived person performed the harmful act or omission, is irrelevant.'
573
+ - '1. Anyone who, by knowingly presenting false facts as true or by unlawfully concealing
574
+ or withholding true facts, damages another person''s property by persuading someone
575
+ to act, omission, or tolerance with the aim of obtaining, for themselves or another,
576
+ an unlawful financial gain from the damage to that property shall be punished
577
+ with imprisonment, "and if the damage caused is particularly great, with imprisonment
578
+ of at least three (3) months and a fine." .
579
+
580
+ If the damage caused exceeds a total of one hundred and twenty thousand (120,000)
581
+ euros, imprisonment of up to ten (10) years and a fine shall be imposed.
582
+
583
+ 2. If the fraud is directed directly against the legal entity of the Greek State,
584
+ legal entities governed by public law, or local government organizations, and
585
+ the damage caused exceeds a total of one hundred and twenty thousand (120,000)
586
+ euros, a prison sentence of at least ten (10) years and a fine of up to one thousand
587
+ (1,000) daily units shall be imposed. This offense shall be time-barred after
588
+ twenty (20) years.
589
+
590
+ '
591
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
592
+
593
+
594
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
595
+ benefit, causes damage to another’s property by persuading someone to act, omit,
596
+ or tolerate something through the knowing misrepresentation of false facts as
597
+ true, or through the unlawful concealment or suppression of true facts, shall
598
+ be punished by imprisonment of at least three months, and if the damage caused
599
+ is particularly large, by imprisonment of at least two years."
600
+
601
+
602
+ From this provision it follows that, for the crime of fraud to be established,
603
+ the following elements are required:
604
+
605
+
606
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
607
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
608
+
609
+
610
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
611
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
612
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
613
+ to themselves or another; and
614
+
615
+
616
+ c) Damage to another person’s property, as defined under civil law, which must
617
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
618
+ not required that the person deceived and the person who suffered the damage be
619
+ the same individual.
620
+
621
+
622
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
623
+ relating to the past or present, and not to those that will occur in the future,
624
+ such as mere promises or contractual obligations. However, when such promises
625
+ or obligations are accompanied by false assurances and representations of other
626
+ false facts referring to the present or the past, in such a manner as to create
627
+ the impression of future fulfillment based on a false present situation fabricated
628
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
629
+ the crime of fraud is established.
630
+
631
+
632
+ The term “property” refers to the totality of a person’s economic assets that
633
+ possess monetary value, while damage to property means its reduction—specifically,
634
+ the difference between the monetary value the property had before the disposition
635
+ caused by the fraudulent conduct and the value remaining after it. Property damage
636
+ exists even if the victim possesses an active claim for restitution.
637
+
638
+
639
+ The time of commission of the fraud is considered to be the moment when the perpetrator
640
+ acted and completed their fraudulent conduct, namely when they made the false
641
+ representations that deceived the victim or a third party. Any subsequent moment
642
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
643
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
644
+ - source_sentence: When is the time of commission of the fraud considered?
645
+ sentences:
646
+ - 'Spear phishing targets specific individuals or employees within an organization
647
+ using personalized, deceptive emails. Unlike mass phishing, these emails are crafted
648
+ to seem familiar and urgent.
649
+
650
+
651
+ Scenarios:
652
+
653
+ - CEO Fraud: Attackers impersonate executives to extract financial or sensitive
654
+ data from employees.
655
+
656
+ - Whaling: High-ranking executives are targeted using tailored fraud emails that
657
+ press for immediate action without verification.'
658
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
659
+
660
+
661
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
662
+ benefit, causes damage to another’s property by persuading someone to act, omit,
663
+ or tolerate something through the knowing misrepresentation of false facts as
664
+ true, or through the unlawful concealment or suppression of true facts, shall
665
+ be punished by imprisonment of at least three months, and if the damage caused
666
+ is particularly large, by imprisonment of at least two years."
667
+
668
+
669
+ From this provision it follows that, for the crime of fraud to be established,
670
+ the following elements are required:
671
+
672
+
673
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
674
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
675
+
676
+
677
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
678
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
679
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
680
+ to themselves or another; and
681
+
682
+
683
+ c) Damage to another person’s property, as defined under civil law, which must
684
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
685
+ not required that the person deceived and the person who suffered the damage be
686
+ the same individual.
687
+
688
+
689
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
690
+ relating to the past or present, and not to those that will occur in the future,
691
+ such as mere promises or contractual obligations. However, when such promises
692
+ or obligations are accompanied by false assurances and representations of other
693
+ false facts referring to the present or the past, in such a manner as to create
694
+ the impression of future fulfillment based on a false present situation fabricated
695
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
696
+ the crime of fraud is established.
697
+
698
+
699
+ The term “property” refers to the totality of a person’s economic assets that
700
+ possess monetary value, while damage to property means its reduction—specifically,
701
+ the difference between the monetary value the property had before the disposition
702
+ caused by the fraudulent conduct and the value remaining after it. Property damage
703
+ exists even if the victim possesses an active claim for restitution.
704
+
705
+
706
+ The time of commission of the fraud is considered to be the moment when the perpetrator
707
+ acted and completed their fraudulent conduct, namely when they made the false
708
+ representations that deceived the victim or a third party. Any subsequent moment
709
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
710
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
711
+ - 'According to Article 386 paragraph 1 of the Greek Penal Code,
712
+
713
+
714
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
715
+ benefit, causes damage to another’s property by persuading someone to act, omit,
716
+ or tolerate something through the knowing misrepresentation of false facts as
717
+ true, or through the unlawful concealment or suppression of true facts, shall
718
+ be punished by imprisonment of at least three months, and if the damage caused
719
+ is particularly large, by imprisonment of at least two years."
720
+
721
+
722
+ From these provisions, it follows that, for the crime of fraud to be established,
723
+ the following elements are required:
724
+
725
+
726
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
727
+ pecuniary benefit;
728
+
729
+
730
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
731
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
732
+ is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves
733
+ or another; and
734
+
735
+
736
+ c) Damage to another’s property, as defined under civil law, which must be causally
737
+ connected to the perpetrator’s deceptive acts.
738
+
739
+
740
+ From the above provisions, it is deduced that the crime of fraud is established
741
+ both objectively and subjectively through the knowing misrepresentation of false
742
+ facts as true, or the unlawful concealment or suppression of true ones, by which
743
+ another person is deceived and, as a result, performs an act, omission, or acquiescence
744
+ involving a disposition of property that directly and necessarily causes financial
745
+ damage to the deceived person or another, with the intent that the perpetrator
746
+ or another gain an unlawful benefit. It is irrelevant whether this intended benefit
747
+ was ultimately achieved.
748
+
749
+
750
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
751
+ relating to the past or present, and not to those expected to occur in the future,
752
+ such as mere promises or contractual obligations. The false fact must have existed
753
+ in the past or must be a present circumstance at the time it is asserted, and
754
+ cannot relate to the future.
755
+
756
+
757
+ However, when future circumstances—that is, promises or contractual obligations—are
758
+ accompanied by false assurances and representations of other false facts referring
759
+ to the present or past, in such a way as to create the impression of future fulfillment,
760
+ based on a false present situation or supposed ability of the perpetrator, who
761
+ had already made the decision not to fulfill their obligation, then the crime
762
+ of fraud is established.'
763
+ pipeline_tag: sentence-similarity
764
+ library_name: sentence-transformers
765
+ metrics:
766
+ - cosine_accuracy@1
767
+ - cosine_accuracy@3
768
+ - cosine_accuracy@5
769
+ - cosine_accuracy@10
770
+ - cosine_precision@1
771
+ - cosine_precision@3
772
+ - cosine_precision@5
773
+ - cosine_precision@10
774
+ - cosine_recall@1
775
+ - cosine_recall@3
776
+ - cosine_recall@5
777
+ - cosine_recall@10
778
+ - cosine_ndcg@10
779
+ - cosine_mrr@10
780
+ - cosine_map@100
781
+ model-index:
782
+ - name: multilingual_e5_large Finetuned on Data
783
+ results:
784
+ - task:
785
+ type: information-retrieval
786
+ name: Information Retrieval
787
+ dataset:
788
+ name: dim 1024
789
+ type: dim_1024
790
+ metrics:
791
+ - type: cosine_accuracy@1
792
+ value: 0.5238095238095238
793
+ name: Cosine Accuracy@1
794
+ - type: cosine_accuracy@3
795
+ value: 0.5238095238095238
796
+ name: Cosine Accuracy@3
797
+ - type: cosine_accuracy@5
798
+ value: 0.5714285714285714
799
+ name: Cosine Accuracy@5
800
+ - type: cosine_accuracy@10
801
+ value: 0.6666666666666666
802
+ name: Cosine Accuracy@10
803
+ - type: cosine_precision@1
804
+ value: 0.5238095238095238
805
+ name: Cosine Precision@1
806
+ - type: cosine_precision@3
807
+ value: 0.5079365079365079
808
+ name: Cosine Precision@3
809
+ - type: cosine_precision@5
810
+ value: 0.47619047619047616
811
+ name: Cosine Precision@5
812
+ - type: cosine_precision@10
813
+ value: 0.4476190476190477
814
+ name: Cosine Precision@10
815
+ - type: cosine_recall@1
816
+ value: 0.08933150183150182
817
+ name: Cosine Recall@1
818
+ - type: cosine_recall@3
819
+ value: 0.24418498168498168
820
+ name: Cosine Recall@3
821
+ - type: cosine_recall@5
822
+ value: 0.33951465201465203
823
+ name: Cosine Recall@5
824
+ - type: cosine_recall@10
825
+ value: 0.5401404151404151
826
+ name: Cosine Recall@10
827
+ - type: cosine_ndcg@10
828
+ value: 0.5921167294151266
829
+ name: Cosine Ndcg@10
830
+ - type: cosine_mrr@10
831
+ value: 0.5480725623582765
832
+ name: Cosine Mrr@10
833
+ - type: cosine_map@100
834
+ value: 0.67423207909377
835
+ name: Cosine Map@100
836
+ - task:
837
+ type: information-retrieval
838
+ name: Information Retrieval
839
+ dataset:
840
+ name: dim 768
841
+ type: dim_768
842
+ metrics:
843
+ - type: cosine_accuracy@1
844
+ value: 0.5238095238095238
845
+ name: Cosine Accuracy@1
846
+ - type: cosine_accuracy@3
847
+ value: 0.5238095238095238
848
+ name: Cosine Accuracy@3
849
+ - type: cosine_accuracy@5
850
+ value: 0.5714285714285714
851
+ name: Cosine Accuracy@5
852
+ - type: cosine_accuracy@10
853
+ value: 0.6666666666666666
854
+ name: Cosine Accuracy@10
855
+ - type: cosine_precision@1
856
+ value: 0.5238095238095238
857
+ name: Cosine Precision@1
858
+ - type: cosine_precision@3
859
+ value: 0.5079365079365079
860
+ name: Cosine Precision@3
861
+ - type: cosine_precision@5
862
+ value: 0.47619047619047616
863
+ name: Cosine Precision@5
864
+ - type: cosine_precision@10
865
+ value: 0.4476190476190477
866
+ name: Cosine Precision@10
867
+ - type: cosine_recall@1
868
+ value: 0.08933150183150182
869
+ name: Cosine Recall@1
870
+ - type: cosine_recall@3
871
+ value: 0.24418498168498168
872
+ name: Cosine Recall@3
873
+ - type: cosine_recall@5
874
+ value: 0.33951465201465203
875
+ name: Cosine Recall@5
876
+ - type: cosine_recall@10
877
+ value: 0.5401404151404151
878
+ name: Cosine Recall@10
879
+ - type: cosine_ndcg@10
880
+ value: 0.5921167294151266
881
+ name: Cosine Ndcg@10
882
+ - type: cosine_mrr@10
883
+ value: 0.5480725623582765
884
+ name: Cosine Mrr@10
885
+ - type: cosine_map@100
886
+ value: 0.67423207909377
887
+ name: Cosine Map@100
888
+ - task:
889
+ type: information-retrieval
890
+ name: Information Retrieval
891
+ dataset:
892
+ name: dim 512
893
+ type: dim_512
894
+ metrics:
895
+ - type: cosine_accuracy@1
896
+ value: 0.47619047619047616
897
+ name: Cosine Accuracy@1
898
+ - type: cosine_accuracy@3
899
+ value: 0.47619047619047616
900
+ name: Cosine Accuracy@3
901
+ - type: cosine_accuracy@5
902
+ value: 0.5714285714285714
903
+ name: Cosine Accuracy@5
904
+ - type: cosine_accuracy@10
905
+ value: 0.6190476190476191
906
+ name: Cosine Accuracy@10
907
+ - type: cosine_precision@1
908
+ value: 0.47619047619047616
909
+ name: Cosine Precision@1
910
+ - type: cosine_precision@3
911
+ value: 0.4603174603174603
912
+ name: Cosine Precision@3
913
+ - type: cosine_precision@5
914
+ value: 0.45714285714285713
915
+ name: Cosine Precision@5
916
+ - type: cosine_precision@10
917
+ value: 0.4238095238095239
918
+ name: Cosine Precision@10
919
+ - type: cosine_recall@1
920
+ value: 0.07345848595848595
921
+ name: Cosine Recall@1
922
+ - type: cosine_recall@3
923
+ value: 0.19656593406593406
924
+ name: Cosine Recall@3
925
+ - type: cosine_recall@5
926
+ value: 0.3077686202686203
927
+ name: Cosine Recall@5
928
+ - type: cosine_recall@10
929
+ value: 0.5202991452991453
930
+ name: Cosine Recall@10
931
+ - type: cosine_ndcg@10
932
+ value: 0.5518338753600308
933
+ name: Cosine Ndcg@10
934
+ - type: cosine_mrr@10
935
+ value: 0.5020408163265305
936
+ name: Cosine Mrr@10
937
+ - type: cosine_map@100
938
+ value: 0.6265911712939339
939
+ name: Cosine Map@100
940
+ - task:
941
+ type: information-retrieval
942
+ name: Information Retrieval
943
+ dataset:
944
+ name: dim 256
945
+ type: dim_256
946
+ metrics:
947
+ - type: cosine_accuracy@1
948
+ value: 0.5238095238095238
949
+ name: Cosine Accuracy@1
950
+ - type: cosine_accuracy@3
951
+ value: 0.5238095238095238
952
+ name: Cosine Accuracy@3
953
+ - type: cosine_accuracy@5
954
+ value: 0.5714285714285714
955
+ name: Cosine Accuracy@5
956
+ - type: cosine_accuracy@10
957
+ value: 0.6190476190476191
958
+ name: Cosine Accuracy@10
959
+ - type: cosine_precision@1
960
+ value: 0.5238095238095238
961
+ name: Cosine Precision@1
962
+ - type: cosine_precision@3
963
+ value: 0.5079365079365079
964
+ name: Cosine Precision@3
965
+ - type: cosine_precision@5
966
+ value: 0.49523809523809514
967
+ name: Cosine Precision@5
968
+ - type: cosine_precision@10
969
+ value: 0.4238095238095239
970
+ name: Cosine Precision@10
971
+ - type: cosine_recall@1
972
+ value: 0.0813949938949939
973
+ name: Cosine Recall@1
974
+ - type: cosine_recall@3
975
+ value: 0.22037545787545787
976
+ name: Cosine Recall@3
977
+ - type: cosine_recall@5
978
+ value: 0.33951465201465203
979
+ name: Cosine Recall@5
980
+ - type: cosine_recall@10
981
+ value: 0.5202991452991453
982
+ name: Cosine Recall@10
983
+ - type: cosine_ndcg@10
984
+ value: 0.5708936958722651
985
+ name: Cosine Ndcg@10
986
+ - type: cosine_mrr@10
987
+ value: 0.5401360544217686
988
+ name: Cosine Mrr@10
989
+ - type: cosine_map@100
990
+ value: 0.651530364911684
991
+ name: Cosine Map@100
992
+ - task:
993
+ type: information-retrieval
994
+ name: Information Retrieval
995
+ dataset:
996
+ name: dim 128
997
+ type: dim_128
998
+ metrics:
999
+ - type: cosine_accuracy@1
1000
+ value: 0.5238095238095238
1001
+ name: Cosine Accuracy@1
1002
+ - type: cosine_accuracy@3
1003
+ value: 0.5238095238095238
1004
+ name: Cosine Accuracy@3
1005
+ - type: cosine_accuracy@5
1006
+ value: 0.5714285714285714
1007
+ name: Cosine Accuracy@5
1008
+ - type: cosine_accuracy@10
1009
+ value: 0.6190476190476191
1010
+ name: Cosine Accuracy@10
1011
+ - type: cosine_precision@1
1012
+ value: 0.5238095238095238
1013
+ name: Cosine Precision@1
1014
+ - type: cosine_precision@3
1015
+ value: 0.5238095238095238
1016
+ name: Cosine Precision@3
1017
+ - type: cosine_precision@5
1018
+ value: 0.5047619047619047
1019
+ name: Cosine Precision@5
1020
+ - type: cosine_precision@10
1021
+ value: 0.4238095238095239
1022
+ name: Cosine Precision@10
1023
+ - type: cosine_recall@1
1024
+ value: 0.07345848595848595
1025
+ name: Cosine Recall@1
1026
+ - type: cosine_recall@3
1027
+ value: 0.2203754578754579
1028
+ name: Cosine Recall@3
1029
+ - type: cosine_recall@5
1030
+ value: 0.34745115995116
1031
+ name: Cosine Recall@5
1032
+ - type: cosine_recall@10
1033
+ value: 0.5202991452991453
1034
+ name: Cosine Recall@10
1035
+ - type: cosine_ndcg@10
1036
+ value: 0.5685354415901852
1037
+ name: Cosine Ndcg@10
1038
+ - type: cosine_mrr@10
1039
+ value: 0.5401360544217686
1040
+ name: Cosine Mrr@10
1041
+ - type: cosine_map@100
1042
+ value: 0.6489604480560528
1043
+ name: Cosine Map@100
1044
+ - task:
1045
+ type: information-retrieval
1046
+ name: Information Retrieval
1047
+ dataset:
1048
+ name: dim 64
1049
+ type: dim_64
1050
+ metrics:
1051
+ - type: cosine_accuracy@1
1052
+ value: 0.42857142857142855
1053
+ name: Cosine Accuracy@1
1054
+ - type: cosine_accuracy@3
1055
+ value: 0.42857142857142855
1056
+ name: Cosine Accuracy@3
1057
+ - type: cosine_accuracy@5
1058
+ value: 0.47619047619047616
1059
+ name: Cosine Accuracy@5
1060
+ - type: cosine_accuracy@10
1061
+ value: 0.6190476190476191
1062
+ name: Cosine Accuracy@10
1063
+ - type: cosine_precision@1
1064
+ value: 0.42857142857142855
1065
+ name: Cosine Precision@1
1066
+ - type: cosine_precision@3
1067
+ value: 0.42857142857142855
1068
+ name: Cosine Precision@3
1069
+ - type: cosine_precision@5
1070
+ value: 0.42857142857142855
1071
+ name: Cosine Precision@5
1072
+ - type: cosine_precision@10
1073
+ value: 0.3999999999999999
1074
+ name: Cosine Precision@10
1075
+ - type: cosine_recall@1
1076
+ value: 0.053617216117216114
1077
+ name: Cosine Recall@1
1078
+ - type: cosine_recall@3
1079
+ value: 0.16085164835164836
1080
+ name: Cosine Recall@3
1081
+ - type: cosine_recall@5
1082
+ value: 0.27205433455433453
1083
+ name: Cosine Recall@5
1084
+ - type: cosine_recall@10
1085
+ value: 0.5004578754578755
1086
+ name: Cosine Recall@10
1087
+ - type: cosine_ndcg@10
1088
+ value: 0.51131642091388
1089
+ name: Cosine Ndcg@10
1090
+ - type: cosine_mrr@10
1091
+ value: 0.45963718820861665
1092
+ name: Cosine Mrr@10
1093
+ - type: cosine_map@100
1094
+ value: 0.5888462989137369
1095
+ name: Cosine Map@100
1096
+ ---
1097
+
1098
+ # multilingual_e5_large Finetuned on Data
1099
+
1100
+ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large). 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.
1101
+
1102
+ ## Model Details
1103
+
1104
+ ### Model Description
1105
+ - **Model Type:** Sentence Transformer
1106
+ - **Base model:** [intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large) <!-- at revision 0dc5580a448e4284468b8909bae50fa925907bc5 -->
1107
+ - **Maximum Sequence Length:** 512 tokens
1108
+ - **Output Dimensionality:** 1024 dimensions
1109
+ - **Similarity Function:** Cosine Similarity
1110
+ <!-- - **Training Dataset:** Unknown -->
1111
+ - **Language:** en
1112
+ - **License:** apache-2.0
1113
+
1114
+ ### Model Sources
1115
+
1116
+ - **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
1117
+ - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
1118
+ - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
1119
+
1120
+ ### Full Model Architecture
1121
+
1122
+ ```
1123
+ SentenceTransformer(
1124
+ (0): Transformer({'max_seq_length': 512, 'do_lower_case': False, 'architecture': 'XLMRobertaModel'})
1125
+ (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
1126
+ (2): Normalize()
1127
+ )
1128
+ ```
1129
+
1130
+ ## Usage
1131
+
1132
+ ### Direct Usage (Sentence Transformers)
1133
+
1134
+ First install the Sentence Transformers library:
1135
+
1136
+ ```bash
1137
+ pip install -U sentence-transformers
1138
+ ```
1139
+
1140
+ Then you can load this model and run inference.
1141
+ ```python
1142
+ from sentence_transformers import SentenceTransformer
1143
+
1144
+ # Download from the 🤗 Hub
1145
+ model = SentenceTransformer("sentence_transformers_model_id")
1146
+ # Run inference
1147
+ sentences = [
1148
+ 'When is the time of commission of the fraud considered?',
1149
+ 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom this provision it follows that, for the crime of fraud to be established, the following elements are required:\n\na) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, without it being necessary that the benefit actually materialize;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence that is detrimental to themselves or another; and\n\nc) Damage to another person’s property, as defined under civil law, which must be causally linked to the deceptive acts or omissions of the perpetrator. It is not required that the person deceived and the person who suffered the damage be the same individual.\n\nThe term “facts”, within the meaning of the above provision, refers to real circumstances relating to the past or present, and not to those that will occur in the future, such as mere promises or contractual obligations. However, when such promises or obligations are accompanied by false assurances and representations of other false facts referring to the present or the past, in such a manner as to create the impression of future fulfillment based on a false present situation fabricated by the perpetrator, who has already formed the decision not to fulfill their obligation, the crime of fraud is established.\n\nThe term “property” refers to the totality of a person’s economic assets that possess monetary value, while damage to property means its reduction—specifically, the difference between the monetary value the property had before the disposition caused by the fraudulent conduct and the value remaining after it. Property damage exists even if the victim possesses an active claim for restitution.\n\nThe time of commission of the fraud is considered to be the moment when the perpetrator acted and completed their fraudulent conduct, namely when they made the false representations that deceived the victim or a third party. Any subsequent moment at which the victim’s damage actually occurred—thereby completing the fraud—or the time when the victim carried out the harmful act or omission, is irrelevant.',
1150
+ 'Spear phishing targets specific individuals or employees within an organization using personalized, deceptive emails. Unlike mass phishing, these emails are crafted to seem familiar and urgent.\n\nScenarios:\n- CEO Fraud: Attackers impersonate executives to extract financial or sensitive data from employees.\n- Whaling: High-ranking executives are targeted using tailored fraud emails that press for immediate action without verification.',
1151
+ ]
1152
+ embeddings = model.encode(sentences)
1153
+ print(embeddings.shape)
1154
+ # [3, 1024]
1155
+
1156
+ # Get the similarity scores for the embeddings
1157
+ similarities = model.similarity(embeddings, embeddings)
1158
+ print(similarities)
1159
+ # tensor([[1.0000, 0.6673, 0.4780],
1160
+ # [0.6673, 1.0000, 0.4691],
1161
+ # [0.4780, 0.4691, 1.0000]])
1162
+ ```
1163
+
1164
+ <!--
1165
+ ### Direct Usage (Transformers)
1166
+
1167
+ <details><summary>Click to see the direct usage in Transformers</summary>
1168
+
1169
+ </details>
1170
+ -->
1171
+
1172
+ <!--
1173
+ ### Downstream Usage (Sentence Transformers)
1174
+
1175
+ You can finetune this model on your own dataset.
1176
+
1177
+ <details><summary>Click to expand</summary>
1178
+
1179
+ </details>
1180
+ -->
1181
+
1182
+ <!--
1183
+ ### Out-of-Scope Use
1184
+
1185
+ *List how the model may foreseeably be misused and address what users ought not to do with the model.*
1186
+ -->
1187
+
1188
+ ## Evaluation
1189
+
1190
+ ### Metrics
1191
+
1192
+ #### Information Retrieval
1193
+
1194
+ * Dataset: `dim_1024`
1195
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1196
+ ```json
1197
+ {
1198
+ "truncate_dim": 1024
1199
+ }
1200
+ ```
1201
+
1202
+ | Metric | Value |
1203
+ |:--------------------|:-----------|
1204
+ | cosine_accuracy@1 | 0.5238 |
1205
+ | cosine_accuracy@3 | 0.5238 |
1206
+ | cosine_accuracy@5 | 0.5714 |
1207
+ | cosine_accuracy@10 | 0.6667 |
1208
+ | cosine_precision@1 | 0.5238 |
1209
+ | cosine_precision@3 | 0.5079 |
1210
+ | cosine_precision@5 | 0.4762 |
1211
+ | cosine_precision@10 | 0.4476 |
1212
+ | cosine_recall@1 | 0.0893 |
1213
+ | cosine_recall@3 | 0.2442 |
1214
+ | cosine_recall@5 | 0.3395 |
1215
+ | cosine_recall@10 | 0.5401 |
1216
+ | **cosine_ndcg@10** | **0.5921** |
1217
+ | cosine_mrr@10 | 0.5481 |
1218
+ | cosine_map@100 | 0.6742 |
1219
+
1220
+ #### Information Retrieval
1221
+
1222
+ * Dataset: `dim_768`
1223
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1224
+ ```json
1225
+ {
1226
+ "truncate_dim": 768
1227
+ }
1228
+ ```
1229
+
1230
+ | Metric | Value |
1231
+ |:--------------------|:-----------|
1232
+ | cosine_accuracy@1 | 0.5238 |
1233
+ | cosine_accuracy@3 | 0.5238 |
1234
+ | cosine_accuracy@5 | 0.5714 |
1235
+ | cosine_accuracy@10 | 0.6667 |
1236
+ | cosine_precision@1 | 0.5238 |
1237
+ | cosine_precision@3 | 0.5079 |
1238
+ | cosine_precision@5 | 0.4762 |
1239
+ | cosine_precision@10 | 0.4476 |
1240
+ | cosine_recall@1 | 0.0893 |
1241
+ | cosine_recall@3 | 0.2442 |
1242
+ | cosine_recall@5 | 0.3395 |
1243
+ | cosine_recall@10 | 0.5401 |
1244
+ | **cosine_ndcg@10** | **0.5921** |
1245
+ | cosine_mrr@10 | 0.5481 |
1246
+ | cosine_map@100 | 0.6742 |
1247
+
1248
+ #### Information Retrieval
1249
+
1250
+ * Dataset: `dim_512`
1251
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1252
+ ```json
1253
+ {
1254
+ "truncate_dim": 512
1255
+ }
1256
+ ```
1257
+
1258
+ | Metric | Value |
1259
+ |:--------------------|:-----------|
1260
+ | cosine_accuracy@1 | 0.4762 |
1261
+ | cosine_accuracy@3 | 0.4762 |
1262
+ | cosine_accuracy@5 | 0.5714 |
1263
+ | cosine_accuracy@10 | 0.619 |
1264
+ | cosine_precision@1 | 0.4762 |
1265
+ | cosine_precision@3 | 0.4603 |
1266
+ | cosine_precision@5 | 0.4571 |
1267
+ | cosine_precision@10 | 0.4238 |
1268
+ | cosine_recall@1 | 0.0735 |
1269
+ | cosine_recall@3 | 0.1966 |
1270
+ | cosine_recall@5 | 0.3078 |
1271
+ | cosine_recall@10 | 0.5203 |
1272
+ | **cosine_ndcg@10** | **0.5518** |
1273
+ | cosine_mrr@10 | 0.502 |
1274
+ | cosine_map@100 | 0.6266 |
1275
+
1276
+ #### Information Retrieval
1277
+
1278
+ * Dataset: `dim_256`
1279
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1280
+ ```json
1281
+ {
1282
+ "truncate_dim": 256
1283
+ }
1284
+ ```
1285
+
1286
+ | Metric | Value |
1287
+ |:--------------------|:-----------|
1288
+ | cosine_accuracy@1 | 0.5238 |
1289
+ | cosine_accuracy@3 | 0.5238 |
1290
+ | cosine_accuracy@5 | 0.5714 |
1291
+ | cosine_accuracy@10 | 0.619 |
1292
+ | cosine_precision@1 | 0.5238 |
1293
+ | cosine_precision@3 | 0.5079 |
1294
+ | cosine_precision@5 | 0.4952 |
1295
+ | cosine_precision@10 | 0.4238 |
1296
+ | cosine_recall@1 | 0.0814 |
1297
+ | cosine_recall@3 | 0.2204 |
1298
+ | cosine_recall@5 | 0.3395 |
1299
+ | cosine_recall@10 | 0.5203 |
1300
+ | **cosine_ndcg@10** | **0.5709** |
1301
+ | cosine_mrr@10 | 0.5401 |
1302
+ | cosine_map@100 | 0.6515 |
1303
+
1304
+ #### Information Retrieval
1305
+
1306
+ * Dataset: `dim_128`
1307
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1308
+ ```json
1309
+ {
1310
+ "truncate_dim": 128
1311
+ }
1312
+ ```
1313
+
1314
+ | Metric | Value |
1315
+ |:--------------------|:-----------|
1316
+ | cosine_accuracy@1 | 0.5238 |
1317
+ | cosine_accuracy@3 | 0.5238 |
1318
+ | cosine_accuracy@5 | 0.5714 |
1319
+ | cosine_accuracy@10 | 0.619 |
1320
+ | cosine_precision@1 | 0.5238 |
1321
+ | cosine_precision@3 | 0.5238 |
1322
+ | cosine_precision@5 | 0.5048 |
1323
+ | cosine_precision@10 | 0.4238 |
1324
+ | cosine_recall@1 | 0.0735 |
1325
+ | cosine_recall@3 | 0.2204 |
1326
+ | cosine_recall@5 | 0.3475 |
1327
+ | cosine_recall@10 | 0.5203 |
1328
+ | **cosine_ndcg@10** | **0.5685** |
1329
+ | cosine_mrr@10 | 0.5401 |
1330
+ | cosine_map@100 | 0.649 |
1331
+
1332
+ #### Information Retrieval
1333
+
1334
+ * Dataset: `dim_64`
1335
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1336
+ ```json
1337
+ {
1338
+ "truncate_dim": 64
1339
+ }
1340
+ ```
1341
+
1342
+ | Metric | Value |
1343
+ |:--------------------|:-----------|
1344
+ | cosine_accuracy@1 | 0.4286 |
1345
+ | cosine_accuracy@3 | 0.4286 |
1346
+ | cosine_accuracy@5 | 0.4762 |
1347
+ | cosine_accuracy@10 | 0.619 |
1348
+ | cosine_precision@1 | 0.4286 |
1349
+ | cosine_precision@3 | 0.4286 |
1350
+ | cosine_precision@5 | 0.4286 |
1351
+ | cosine_precision@10 | 0.4 |
1352
+ | cosine_recall@1 | 0.0536 |
1353
+ | cosine_recall@3 | 0.1609 |
1354
+ | cosine_recall@5 | 0.2721 |
1355
+ | cosine_recall@10 | 0.5005 |
1356
+ | **cosine_ndcg@10** | **0.5113** |
1357
+ | cosine_mrr@10 | 0.4596 |
1358
+ | cosine_map@100 | 0.5888 |
1359
+
1360
+ <!--
1361
+ ## Bias, Risks and Limitations
1362
+
1363
+ *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
1364
+ -->
1365
+
1366
+ <!--
1367
+ ### Recommendations
1368
+
1369
+ *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
1370
+ -->
1371
+
1372
+ ## Training Details
1373
+
1374
+ ### Training Dataset
1375
+
1376
+ #### Unnamed Dataset
1377
+
1378
+ * Size: 82 training samples
1379
+ * Columns: <code>anchor</code> and <code>positive</code>
1380
+ * Approximate statistics based on the first 82 samples:
1381
+ | | anchor | positive |
1382
+ |:--------|:----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|
1383
+ | type | string | string |
1384
+ | details | <ul><li>min: 9 tokens</li><li>mean: 18.17 tokens</li><li>max: 34 tokens</li></ul> | <ul><li>min: 69 tokens</li><li>mean: 399.51 tokens</li><li>max: 512 tokens</li></ul> |
1385
+ * Samples:
1386
+ | anchor | positive |
1387
+ |:----------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
1388
+ | <code>What determines whether the act in question shall be punished if the offender is in the service of the legal holder of the data?</code> | <code>Everyone who obtains access to data recorded in a computer or in the external memory of a computer or transmitted by telecommunication systems shall be punished with imprisonment for up to six months or by a fine from 29 to 15,000 Euro, under the condition that these acts have been committed without right, especially in violation of prohibitions or of security measures taken by the legal holder. If the act concerns the international relations or the security of the State, he shall be punished according to Article 148.<br>If the offender is in the service of the legal holder of the data, the act of the preceding paragraph shall be punished only if it has been explicitly prohibited by internal regulations or by a written decision of the holder or of a competent employee of his.<br></code> |
1389
+ | <code>What must be causally connected to the perpetrator's deceptive acts?</code> | <code>According to Article 386 paragraph 1 of the Greek Penal Code,<br><br>"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."<br><br>From these provisions, it follows that, for the crime of fraud to be established, the following elements are required:<br><br>a) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit;<br><br>b) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence detrimental to th...</code> |
1390
+ | <code>Who can be punished with imprisonment?</code> | <code>1. Anyone who, by knowingly presenting false facts as true or by unlawfully concealing or withholding true facts, damages another person's property by persuading someone to act, omission, or tolerance with the aim of obtaining, for themselves or another, an unlawful financial gain from the damage to that property shall be punished with imprisonment, "and if the damage caused is particularly great, with imprisonment of at least three (3) months and a fine." .<br>If the damage caused exceeds a total of one hundred and twenty thousand (120,000) euros, imprisonment of up to ten (10) years and a fine shall be imposed.<br>2. If the fraud is directed directly against the legal entity of the Greek State, legal entities governed by public law, or local government organizations, and the damage caused exceeds a total of one hundred and twenty thousand (120,000) euros, a prison sentence of at least ten (10) years and a fine of up to one thousand (1,000) daily units shall be imposed. This offense shall b...</code> |
1391
+ * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters:
1392
+ ```json
1393
+ {
1394
+ "loss": "MultipleNegativesRankingLoss",
1395
+ "matryoshka_dims": [
1396
+ 1024,
1397
+ 768,
1398
+ 512,
1399
+ 256,
1400
+ 128,
1401
+ 64
1402
+ ],
1403
+ "matryoshka_weights": [
1404
+ 1,
1405
+ 1,
1406
+ 1,
1407
+ 1,
1408
+ 1,
1409
+ 1
1410
+ ],
1411
+ "n_dims_per_step": -1
1412
+ }
1413
+ ```
1414
+
1415
+ ### Training Hyperparameters
1416
+ #### Non-Default Hyperparameters
1417
+
1418
+ - `eval_strategy`: epoch
1419
+ - `gradient_accumulation_steps`: 2
1420
+ - `learning_rate`: 2e-05
1421
+ - `num_train_epochs`: 10
1422
+ - `lr_scheduler_type`: cosine
1423
+ - `warmup_ratio`: 0.1
1424
+ - `bf16`: True
1425
+ - `tf32`: True
1426
+ - `load_best_model_at_end`: True
1427
+ - `optim`: adamw_torch_fused
1428
+ - `batch_sampler`: no_duplicates
1429
+
1430
+ #### All Hyperparameters
1431
+ <details><summary>Click to expand</summary>
1432
+
1433
+ - `overwrite_output_dir`: False
1434
+ - `do_predict`: False
1435
+ - `eval_strategy`: epoch
1436
+ - `prediction_loss_only`: True
1437
+ - `per_device_train_batch_size`: 8
1438
+ - `per_device_eval_batch_size`: 8
1439
+ - `per_gpu_train_batch_size`: None
1440
+ - `per_gpu_eval_batch_size`: None
1441
+ - `gradient_accumulation_steps`: 2
1442
+ - `eval_accumulation_steps`: None
1443
+ - `torch_empty_cache_steps`: None
1444
+ - `learning_rate`: 2e-05
1445
+ - `weight_decay`: 0.0
1446
+ - `adam_beta1`: 0.9
1447
+ - `adam_beta2`: 0.999
1448
+ - `adam_epsilon`: 1e-08
1449
+ - `max_grad_norm`: 1.0
1450
+ - `num_train_epochs`: 10
1451
+ - `max_steps`: -1
1452
+ - `lr_scheduler_type`: cosine
1453
+ - `lr_scheduler_kwargs`: {}
1454
+ - `warmup_ratio`: 0.1
1455
+ - `warmup_steps`: 0
1456
+ - `log_level`: passive
1457
+ - `log_level_replica`: warning
1458
+ - `log_on_each_node`: True
1459
+ - `logging_nan_inf_filter`: True
1460
+ - `save_safetensors`: True
1461
+ - `save_on_each_node`: False
1462
+ - `save_only_model`: False
1463
+ - `restore_callback_states_from_checkpoint`: False
1464
+ - `no_cuda`: False
1465
+ - `use_cpu`: False
1466
+ - `use_mps_device`: False
1467
+ - `seed`: 42
1468
+ - `data_seed`: None
1469
+ - `jit_mode_eval`: False
1470
+ - `use_ipex`: False
1471
+ - `bf16`: True
1472
+ - `fp16`: False
1473
+ - `fp16_opt_level`: O1
1474
+ - `half_precision_backend`: auto
1475
+ - `bf16_full_eval`: False
1476
+ - `fp16_full_eval`: False
1477
+ - `tf32`: True
1478
+ - `local_rank`: 0
1479
+ - `ddp_backend`: None
1480
+ - `tpu_num_cores`: None
1481
+ - `tpu_metrics_debug`: False
1482
+ - `debug`: []
1483
+ - `dataloader_drop_last`: False
1484
+ - `dataloader_num_workers`: 0
1485
+ - `dataloader_prefetch_factor`: None
1486
+ - `past_index`: -1
1487
+ - `disable_tqdm`: False
1488
+ - `remove_unused_columns`: True
1489
+ - `label_names`: None
1490
+ - `load_best_model_at_end`: True
1491
+ - `ignore_data_skip`: False
1492
+ - `fsdp`: []
1493
+ - `fsdp_min_num_params`: 0
1494
+ - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
1495
+ - `tp_size`: 0
1496
+ - `fsdp_transformer_layer_cls_to_wrap`: None
1497
+ - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
1498
+ - `deepspeed`: None
1499
+ - `label_smoothing_factor`: 0.0
1500
+ - `optim`: adamw_torch_fused
1501
+ - `optim_args`: None
1502
+ - `adafactor`: False
1503
+ - `group_by_length`: False
1504
+ - `length_column_name`: length
1505
+ - `ddp_find_unused_parameters`: None
1506
+ - `ddp_bucket_cap_mb`: None
1507
+ - `ddp_broadcast_buffers`: False
1508
+ - `dataloader_pin_memory`: True
1509
+ - `dataloader_persistent_workers`: False
1510
+ - `skip_memory_metrics`: True
1511
+ - `use_legacy_prediction_loop`: False
1512
+ - `push_to_hub`: False
1513
+ - `resume_from_checkpoint`: None
1514
+ - `hub_model_id`: None
1515
+ - `hub_strategy`: every_save
1516
+ - `hub_private_repo`: None
1517
+ - `hub_always_push`: False
1518
+ - `gradient_checkpointing`: False
1519
+ - `gradient_checkpointing_kwargs`: None
1520
+ - `include_inputs_for_metrics`: False
1521
+ - `include_for_metrics`: []
1522
+ - `eval_do_concat_batches`: True
1523
+ - `fp16_backend`: auto
1524
+ - `push_to_hub_model_id`: None
1525
+ - `push_to_hub_organization`: None
1526
+ - `mp_parameters`:
1527
+ - `auto_find_batch_size`: False
1528
+ - `full_determinism`: False
1529
+ - `torchdynamo`: None
1530
+ - `ray_scope`: last
1531
+ - `ddp_timeout`: 1800
1532
+ - `torch_compile`: False
1533
+ - `torch_compile_backend`: None
1534
+ - `torch_compile_mode`: None
1535
+ - `include_tokens_per_second`: False
1536
+ - `include_num_input_tokens_seen`: False
1537
+ - `neftune_noise_alpha`: None
1538
+ - `optim_target_modules`: None
1539
+ - `batch_eval_metrics`: False
1540
+ - `eval_on_start`: False
1541
+ - `use_liger_kernel`: False
1542
+ - `eval_use_gather_object`: False
1543
+ - `average_tokens_across_devices`: False
1544
+ - `prompts`: None
1545
+ - `batch_sampler`: no_duplicates
1546
+ - `multi_dataset_batch_sampler`: proportional
1547
+ - `router_mapping`: {}
1548
+ - `learning_rate_mapping`: {}
1549
+
1550
+ </details>
1551
+
1552
+ ### Training Logs
1553
+ | Epoch | Step | Training Loss | dim_1024_cosine_ndcg@10 | dim_768_cosine_ndcg@10 | dim_512_cosine_ndcg@10 | dim_256_cosine_ndcg@10 | dim_128_cosine_ndcg@10 | dim_64_cosine_ndcg@10 |
1554
+ |:------:|:----:|:-------------:|:-----------------------:|:----------------------:|:----------------------:|:----------------------:|:----------------------:|:---------------------:|
1555
+ | 0.1818 | 1 | 18.029 | - | - | - | - | - | - |
1556
+ | 0.3636 | 2 | 19.4106 | - | - | - | - | - | - |
1557
+ | 0.5455 | 3 | 16.6201 | - | - | - | - | - | - |
1558
+ | 0.7273 | 4 | 15.3048 | - | - | - | - | - | - |
1559
+ | 0.9091 | 5 | 14.0182 | - | - | - | - | - | - |
1560
+ | 1.0 | 6 | 6.4771 | - | - | - | - | - | - |
1561
+ | 1.0909 | 7 | 6.7664 | 0.6167 | 0.5821 | 0.5524 | 0.5177 | 0.5278 | 0.4124 |
1562
+ | 1.1818 | 8 | 11.8583 | - | - | - | - | - | - |
1563
+ | 1.3636 | 9 | 11.9216 | - | - | - | - | - | - |
1564
+ | 1.5455 | 10 | 13.3764 | - | - | - | - | - | - |
1565
+ | 1.7273 | 11 | 12.9063 | - | - | - | - | - | - |
1566
+ | 1.9091 | 12 | 13.5984 | - | - | - | - | - | - |
1567
+ | 2.0 | 13 | 7.8523 | - | - | - | - | - | - |
1568
+ | 2.0909 | 14 | 4.4487 | 0.5921 | 0.5921 | 0.5518 | 0.5709 | 0.5685 | 0.5113 |
1569
+
1570
+
1571
+ ### Framework Versions
1572
+ - Python: 3.12.12
1573
+ - Sentence Transformers: 5.1.1
1574
+ - Transformers: 4.51.3
1575
+ - PyTorch: 2.8.0+cu126
1576
+ - Accelerate: 1.11.0
1577
+ - Datasets: 4.0.0
1578
+ - Tokenizers: 0.21.4
1579
+
1580
+ ## Citation
1581
+
1582
+ ### BibTeX
1583
+
1584
+ #### Sentence Transformers
1585
+ ```bibtex
1586
+ @inproceedings{reimers-2019-sentence-bert,
1587
+ title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
1588
+ author = "Reimers, Nils and Gurevych, Iryna",
1589
+ booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
1590
+ month = "11",
1591
+ year = "2019",
1592
+ publisher = "Association for Computational Linguistics",
1593
+ url = "https://arxiv.org/abs/1908.10084",
1594
+ }
1595
+ ```
1596
+
1597
+ #### MatryoshkaLoss
1598
+ ```bibtex
1599
+ @misc{kusupati2024matryoshka,
1600
+ title={Matryoshka Representation Learning},
1601
+ 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},
1602
+ year={2024},
1603
+ eprint={2205.13147},
1604
+ archivePrefix={arXiv},
1605
+ primaryClass={cs.LG}
1606
+ }
1607
+ ```
1608
+
1609
+ #### MultipleNegativesRankingLoss
1610
+ ```bibtex
1611
+ @misc{henderson2017efficient,
1612
+ title={Efficient Natural Language Response Suggestion for Smart Reply},
1613
+ 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},
1614
+ year={2017},
1615
+ eprint={1705.00652},
1616
+ archivePrefix={arXiv},
1617
+ primaryClass={cs.CL}
1618
+ }
1619
+ ```
1620
+
1621
+ <!--
1622
+ ## Glossary
1623
+
1624
+ *Clearly define terms in order to be accessible across audiences.*
1625
+ -->
1626
+
1627
+ <!--
1628
+ ## Model Card Authors
1629
+
1630
+ *Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
1631
+ -->
1632
+
1633
+ <!--
1634
+ ## Model Card Contact
1635
+
1636
+ *Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
1637
+ -->
checkpoint-14/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "XLMRobertaModel"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 0,
7
+ "classifier_dropout": null,
8
+ "eos_token_id": 2,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 1024,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 4096,
14
+ "layer_norm_eps": 1e-05,
15
+ "max_position_embeddings": 514,
16
+ "model_type": "xlm-roberta",
17
+ "num_attention_heads": 16,
18
+ "num_hidden_layers": 24,
19
+ "output_past": true,
20
+ "pad_token_id": 1,
21
+ "position_embedding_type": "absolute",
22
+ "torch_dtype": "float32",
23
+ "transformers_version": "4.51.3",
24
+ "type_vocab_size": 1,
25
+ "use_cache": true,
26
+ "vocab_size": 250002
27
+ }
checkpoint-14/config_sentence_transformers.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "SentenceTransformer",
3
+ "__version__": {
4
+ "sentence_transformers": "5.1.1",
5
+ "transformers": "4.51.3",
6
+ "pytorch": "2.8.0+cu126"
7
+ },
8
+ "prompts": {
9
+ "query": "",
10
+ "document": ""
11
+ },
12
+ "default_prompt_name": null,
13
+ "similarity_fn_name": "cosine"
14
+ }
checkpoint-14/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46e99ac0a0cb9fa54e6b9a6f77368aacd6fac9b3751511f22a5ca5c9fb7c5204
3
+ size 2239607176
checkpoint-14/modules.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "idx": 0,
4
+ "name": "0",
5
+ "path": "",
6
+ "type": "sentence_transformers.models.Transformer"
7
+ },
8
+ {
9
+ "idx": 1,
10
+ "name": "1",
11
+ "path": "1_Pooling",
12
+ "type": "sentence_transformers.models.Pooling"
13
+ },
14
+ {
15
+ "idx": 2,
16
+ "name": "2",
17
+ "path": "2_Normalize",
18
+ "type": "sentence_transformers.models.Normalize"
19
+ }
20
+ ]
checkpoint-14/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12e6297d4d18fe9514b828db9f04e257311c390afeb2ee8199df0c8a21c372ad
3
+ size 4471067142
checkpoint-14/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68b2584240053bafbf0d1841cb2d9fbbccb6afaacb562198511df474996d98d8
3
+ size 14645
checkpoint-14/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62064cbcdbbfd5a9d97ecb6c2c1900ba606dc24883f0bbf53e596e2266d6824a
3
+ size 1465
checkpoint-14/sentence_bert_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "max_seq_length": 512,
3
+ "do_lower_case": false
4
+ }
checkpoint-14/sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
3
+ size 5069051
checkpoint-14/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
checkpoint-14/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:883b037111086fd4dfebbbc9b7cee11e1517b5e0c0514879478661440f137085
3
+ size 17082987
checkpoint-14/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": true,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
checkpoint-14/trainer_state.json ADDED
@@ -0,0 +1,337 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 14,
3
+ "best_metric": 0.5685354415901852,
4
+ "best_model_checkpoint": "multilingual-e5-large/checkpoint-14",
5
+ "epoch": 2.090909090909091,
6
+ "eval_steps": 500,
7
+ "global_step": 14,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.18181818181818182,
14
+ "grad_norm": 135.6220703125,
15
+ "learning_rate": 0.0,
16
+ "loss": 18.029,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.36363636363636365,
21
+ "grad_norm": 140.5317840576172,
22
+ "learning_rate": 4.000000000000001e-06,
23
+ "loss": 19.4106,
24
+ "step": 2
25
+ },
26
+ {
27
+ "epoch": 0.5454545454545454,
28
+ "grad_norm": 142.2399444580078,
29
+ "learning_rate": 8.000000000000001e-06,
30
+ "loss": 16.6201,
31
+ "step": 3
32
+ },
33
+ {
34
+ "epoch": 0.7272727272727273,
35
+ "grad_norm": 110.63945770263672,
36
+ "learning_rate": 1.2e-05,
37
+ "loss": 15.3048,
38
+ "step": 4
39
+ },
40
+ {
41
+ "epoch": 0.9090909090909091,
42
+ "grad_norm": 129.43089294433594,
43
+ "learning_rate": 1.6000000000000003e-05,
44
+ "loss": 14.0182,
45
+ "step": 5
46
+ },
47
+ {
48
+ "epoch": 1.0,
49
+ "grad_norm": 108.25520324707031,
50
+ "learning_rate": 2e-05,
51
+ "loss": 6.4771,
52
+ "step": 6
53
+ },
54
+ {
55
+ "epoch": 1.0909090909090908,
56
+ "grad_norm": 95.05782318115234,
57
+ "learning_rate": 1.9975640502598243e-05,
58
+ "loss": 6.7664,
59
+ "step": 7
60
+ },
61
+ {
62
+ "epoch": 1.0909090909090908,
63
+ "eval_dim_1024_cosine_accuracy@1": 0.5714285714285714,
64
+ "eval_dim_1024_cosine_accuracy@10": 0.6666666666666666,
65
+ "eval_dim_1024_cosine_accuracy@3": 0.5714285714285714,
66
+ "eval_dim_1024_cosine_accuracy@5": 0.6190476190476191,
67
+ "eval_dim_1024_cosine_map@100": 0.6946498352795571,
68
+ "eval_dim_1024_cosine_mrr@10": 0.5888888888888888,
69
+ "eval_dim_1024_cosine_ndcg@10": 0.6167053425672016,
70
+ "eval_dim_1024_cosine_precision@1": 0.5714285714285714,
71
+ "eval_dim_1024_cosine_precision@10": 0.4476190476190477,
72
+ "eval_dim_1024_cosine_precision@3": 0.5555555555555556,
73
+ "eval_dim_1024_cosine_precision@5": 0.5142857142857142,
74
+ "eval_dim_1024_cosine_recall@1": 0.10123626373626372,
75
+ "eval_dim_1024_cosine_recall@10": 0.5401404151404151,
76
+ "eval_dim_1024_cosine_recall@3": 0.2798992673992674,
77
+ "eval_dim_1024_cosine_recall@5": 0.3871336996336997,
78
+ "eval_dim_128_cosine_accuracy@1": 0.47619047619047616,
79
+ "eval_dim_128_cosine_accuracy@10": 0.5714285714285714,
80
+ "eval_dim_128_cosine_accuracy@3": 0.47619047619047616,
81
+ "eval_dim_128_cosine_accuracy@5": 0.5238095238095238,
82
+ "eval_dim_128_cosine_map@100": 0.5954946117864774,
83
+ "eval_dim_128_cosine_mrr@10": 0.49251700680272104,
84
+ "eval_dim_128_cosine_ndcg@10": 0.5277801377642622,
85
+ "eval_dim_128_cosine_precision@1": 0.47619047619047616,
86
+ "eval_dim_128_cosine_precision@10": 0.3761904761904762,
87
+ "eval_dim_128_cosine_precision@3": 0.4603174603174603,
88
+ "eval_dim_128_cosine_precision@5": 0.42857142857142855,
89
+ "eval_dim_128_cosine_recall@1": 0.08536324786324785,
90
+ "eval_dim_128_cosine_recall@10": 0.4806166056166057,
91
+ "eval_dim_128_cosine_recall@3": 0.23228021978021982,
92
+ "eval_dim_128_cosine_recall@5": 0.31967338217338215,
93
+ "eval_dim_256_cosine_accuracy@1": 0.47619047619047616,
94
+ "eval_dim_256_cosine_accuracy@10": 0.5714285714285714,
95
+ "eval_dim_256_cosine_accuracy@3": 0.47619047619047616,
96
+ "eval_dim_256_cosine_accuracy@5": 0.5238095238095238,
97
+ "eval_dim_256_cosine_map@100": 0.6218563539505393,
98
+ "eval_dim_256_cosine_mrr@10": 0.4925170068027211,
99
+ "eval_dim_256_cosine_ndcg@10": 0.5177339964253599,
100
+ "eval_dim_256_cosine_precision@1": 0.47619047619047616,
101
+ "eval_dim_256_cosine_precision@10": 0.3476190476190476,
102
+ "eval_dim_256_cosine_precision@3": 0.4603174603174603,
103
+ "eval_dim_256_cosine_precision@5": 0.419047619047619,
104
+ "eval_dim_256_cosine_recall@1": 0.09543650793650793,
105
+ "eval_dim_256_cosine_recall@10": 0.47817460317460325,
106
+ "eval_dim_256_cosine_recall@3": 0.2625,
107
+ "eval_dim_256_cosine_recall@5": 0.35813492063492064,
108
+ "eval_dim_512_cosine_accuracy@1": 0.5238095238095238,
109
+ "eval_dim_512_cosine_accuracy@10": 0.5714285714285714,
110
+ "eval_dim_512_cosine_accuracy@3": 0.5238095238095238,
111
+ "eval_dim_512_cosine_accuracy@5": 0.5714285714285714,
112
+ "eval_dim_512_cosine_map@100": 0.6608463504287858,
113
+ "eval_dim_512_cosine_mrr@10": 0.5333333333333333,
114
+ "eval_dim_512_cosine_ndcg@10": 0.5523687509163372,
115
+ "eval_dim_512_cosine_precision@1": 0.5238095238095238,
116
+ "eval_dim_512_cosine_precision@10": 0.3761904761904762,
117
+ "eval_dim_512_cosine_precision@3": 0.5079365079365079,
118
+ "eval_dim_512_cosine_precision@5": 0.4666666666666666,
119
+ "eval_dim_512_cosine_recall@1": 0.0994047619047619,
120
+ "eval_dim_512_cosine_recall@10": 0.5019841269841271,
121
+ "eval_dim_512_cosine_recall@3": 0.2744047619047619,
122
+ "eval_dim_512_cosine_recall@5": 0.37797619047619047,
123
+ "eval_dim_64_cosine_accuracy@1": 0.3333333333333333,
124
+ "eval_dim_64_cosine_accuracy@10": 0.5714285714285714,
125
+ "eval_dim_64_cosine_accuracy@3": 0.38095238095238093,
126
+ "eval_dim_64_cosine_accuracy@5": 0.38095238095238093,
127
+ "eval_dim_64_cosine_map@100": 0.5579595177809107,
128
+ "eval_dim_64_cosine_mrr@10": 0.3764172335600907,
129
+ "eval_dim_64_cosine_ndcg@10": 0.41244392103396355,
130
+ "eval_dim_64_cosine_precision@1": 0.3333333333333333,
131
+ "eval_dim_64_cosine_precision@10": 0.2619047619047619,
132
+ "eval_dim_64_cosine_precision@3": 0.3492063492063492,
133
+ "eval_dim_64_cosine_precision@5": 0.3142857142857143,
134
+ "eval_dim_64_cosine_recall@1": 0.07063492063492063,
135
+ "eval_dim_64_cosine_recall@10": 0.42002442002442003,
136
+ "eval_dim_64_cosine_recall@3": 0.2357142857142857,
137
+ "eval_dim_64_cosine_recall@5": 0.34523809523809523,
138
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
139
+ "eval_dim_768_cosine_accuracy@10": 0.6666666666666666,
140
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
141
+ "eval_dim_768_cosine_accuracy@5": 0.5714285714285714,
142
+ "eval_dim_768_cosine_map@100": 0.6727109022414055,
143
+ "eval_dim_768_cosine_mrr@10": 0.5480725623582765,
144
+ "eval_dim_768_cosine_ndcg@10": 0.5820705880762242,
145
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
146
+ "eval_dim_768_cosine_precision@10": 0.41904761904761906,
147
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
148
+ "eval_dim_768_cosine_precision@5": 0.4666666666666666,
149
+ "eval_dim_768_cosine_recall@1": 0.09726800976800977,
150
+ "eval_dim_768_cosine_recall@10": 0.5163308913308913,
151
+ "eval_dim_768_cosine_recall@3": 0.2679945054945055,
152
+ "eval_dim_768_cosine_recall@5": 0.36729242979242976,
153
+ "eval_runtime": 7.0357,
154
+ "eval_samples_per_second": 0.0,
155
+ "eval_sequential_score": 0.41244392103396355,
156
+ "eval_steps_per_second": 0.0,
157
+ "step": 7
158
+ },
159
+ {
160
+ "epoch": 1.1818181818181819,
161
+ "grad_norm": 159.73828125,
162
+ "learning_rate": 1.9902680687415704e-05,
163
+ "loss": 11.8583,
164
+ "step": 8
165
+ },
166
+ {
167
+ "epoch": 1.3636363636363638,
168
+ "grad_norm": 134.70603942871094,
169
+ "learning_rate": 1.9781476007338058e-05,
170
+ "loss": 11.9216,
171
+ "step": 9
172
+ },
173
+ {
174
+ "epoch": 1.5454545454545454,
175
+ "grad_norm": 139.73654174804688,
176
+ "learning_rate": 1.961261695938319e-05,
177
+ "loss": 13.3764,
178
+ "step": 10
179
+ },
180
+ {
181
+ "epoch": 1.7272727272727273,
182
+ "grad_norm": 173.23065185546875,
183
+ "learning_rate": 1.9396926207859085e-05,
184
+ "loss": 12.9063,
185
+ "step": 11
186
+ },
187
+ {
188
+ "epoch": 1.9090909090909092,
189
+ "grad_norm": 208.78482055664062,
190
+ "learning_rate": 1.913545457642601e-05,
191
+ "loss": 13.5984,
192
+ "step": 12
193
+ },
194
+ {
195
+ "epoch": 2.0,
196
+ "grad_norm": 298.82183837890625,
197
+ "learning_rate": 1.8829475928589272e-05,
198
+ "loss": 7.8523,
199
+ "step": 13
200
+ },
201
+ {
202
+ "epoch": 2.090909090909091,
203
+ "grad_norm": 156.5718994140625,
204
+ "learning_rate": 1.848048096156426e-05,
205
+ "loss": 4.4487,
206
+ "step": 14
207
+ },
208
+ {
209
+ "epoch": 2.090909090909091,
210
+ "eval_dim_1024_cosine_accuracy@1": 0.5238095238095238,
211
+ "eval_dim_1024_cosine_accuracy@10": 0.6666666666666666,
212
+ "eval_dim_1024_cosine_accuracy@3": 0.5238095238095238,
213
+ "eval_dim_1024_cosine_accuracy@5": 0.5714285714285714,
214
+ "eval_dim_1024_cosine_map@100": 0.67423207909377,
215
+ "eval_dim_1024_cosine_mrr@10": 0.5480725623582765,
216
+ "eval_dim_1024_cosine_ndcg@10": 0.5921167294151266,
217
+ "eval_dim_1024_cosine_precision@1": 0.5238095238095238,
218
+ "eval_dim_1024_cosine_precision@10": 0.4476190476190477,
219
+ "eval_dim_1024_cosine_precision@3": 0.5079365079365079,
220
+ "eval_dim_1024_cosine_precision@5": 0.47619047619047616,
221
+ "eval_dim_1024_cosine_recall@1": 0.08933150183150182,
222
+ "eval_dim_1024_cosine_recall@10": 0.5401404151404151,
223
+ "eval_dim_1024_cosine_recall@3": 0.24418498168498168,
224
+ "eval_dim_1024_cosine_recall@5": 0.33951465201465203,
225
+ "eval_dim_128_cosine_accuracy@1": 0.5238095238095238,
226
+ "eval_dim_128_cosine_accuracy@10": 0.6190476190476191,
227
+ "eval_dim_128_cosine_accuracy@3": 0.5238095238095238,
228
+ "eval_dim_128_cosine_accuracy@5": 0.5714285714285714,
229
+ "eval_dim_128_cosine_map@100": 0.6489604480560528,
230
+ "eval_dim_128_cosine_mrr@10": 0.5401360544217686,
231
+ "eval_dim_128_cosine_ndcg@10": 0.5685354415901852,
232
+ "eval_dim_128_cosine_precision@1": 0.5238095238095238,
233
+ "eval_dim_128_cosine_precision@10": 0.4238095238095239,
234
+ "eval_dim_128_cosine_precision@3": 0.5238095238095238,
235
+ "eval_dim_128_cosine_precision@5": 0.5047619047619047,
236
+ "eval_dim_128_cosine_recall@1": 0.07345848595848595,
237
+ "eval_dim_128_cosine_recall@10": 0.5202991452991453,
238
+ "eval_dim_128_cosine_recall@3": 0.2203754578754579,
239
+ "eval_dim_128_cosine_recall@5": 0.34745115995116,
240
+ "eval_dim_256_cosine_accuracy@1": 0.5238095238095238,
241
+ "eval_dim_256_cosine_accuracy@10": 0.6190476190476191,
242
+ "eval_dim_256_cosine_accuracy@3": 0.5238095238095238,
243
+ "eval_dim_256_cosine_accuracy@5": 0.5714285714285714,
244
+ "eval_dim_256_cosine_map@100": 0.651530364911684,
245
+ "eval_dim_256_cosine_mrr@10": 0.5401360544217686,
246
+ "eval_dim_256_cosine_ndcg@10": 0.5708936958722651,
247
+ "eval_dim_256_cosine_precision@1": 0.5238095238095238,
248
+ "eval_dim_256_cosine_precision@10": 0.4238095238095239,
249
+ "eval_dim_256_cosine_precision@3": 0.5079365079365079,
250
+ "eval_dim_256_cosine_precision@5": 0.49523809523809514,
251
+ "eval_dim_256_cosine_recall@1": 0.0813949938949939,
252
+ "eval_dim_256_cosine_recall@10": 0.5202991452991453,
253
+ "eval_dim_256_cosine_recall@3": 0.22037545787545787,
254
+ "eval_dim_256_cosine_recall@5": 0.33951465201465203,
255
+ "eval_dim_512_cosine_accuracy@1": 0.47619047619047616,
256
+ "eval_dim_512_cosine_accuracy@10": 0.6190476190476191,
257
+ "eval_dim_512_cosine_accuracy@3": 0.47619047619047616,
258
+ "eval_dim_512_cosine_accuracy@5": 0.5714285714285714,
259
+ "eval_dim_512_cosine_map@100": 0.6265911712939339,
260
+ "eval_dim_512_cosine_mrr@10": 0.5020408163265305,
261
+ "eval_dim_512_cosine_ndcg@10": 0.5518338753600308,
262
+ "eval_dim_512_cosine_precision@1": 0.47619047619047616,
263
+ "eval_dim_512_cosine_precision@10": 0.4238095238095239,
264
+ "eval_dim_512_cosine_precision@3": 0.4603174603174603,
265
+ "eval_dim_512_cosine_precision@5": 0.45714285714285713,
266
+ "eval_dim_512_cosine_recall@1": 0.07345848595848595,
267
+ "eval_dim_512_cosine_recall@10": 0.5202991452991453,
268
+ "eval_dim_512_cosine_recall@3": 0.19656593406593406,
269
+ "eval_dim_512_cosine_recall@5": 0.3077686202686203,
270
+ "eval_dim_64_cosine_accuracy@1": 0.42857142857142855,
271
+ "eval_dim_64_cosine_accuracy@10": 0.6190476190476191,
272
+ "eval_dim_64_cosine_accuracy@3": 0.42857142857142855,
273
+ "eval_dim_64_cosine_accuracy@5": 0.47619047619047616,
274
+ "eval_dim_64_cosine_map@100": 0.5888462989137369,
275
+ "eval_dim_64_cosine_mrr@10": 0.45963718820861665,
276
+ "eval_dim_64_cosine_ndcg@10": 0.51131642091388,
277
+ "eval_dim_64_cosine_precision@1": 0.42857142857142855,
278
+ "eval_dim_64_cosine_precision@10": 0.3999999999999999,
279
+ "eval_dim_64_cosine_precision@3": 0.42857142857142855,
280
+ "eval_dim_64_cosine_precision@5": 0.42857142857142855,
281
+ "eval_dim_64_cosine_recall@1": 0.053617216117216114,
282
+ "eval_dim_64_cosine_recall@10": 0.5004578754578755,
283
+ "eval_dim_64_cosine_recall@3": 0.16085164835164836,
284
+ "eval_dim_64_cosine_recall@5": 0.27205433455433453,
285
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
286
+ "eval_dim_768_cosine_accuracy@10": 0.6666666666666666,
287
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
288
+ "eval_dim_768_cosine_accuracy@5": 0.5714285714285714,
289
+ "eval_dim_768_cosine_map@100": 0.67423207909377,
290
+ "eval_dim_768_cosine_mrr@10": 0.5480725623582765,
291
+ "eval_dim_768_cosine_ndcg@10": 0.5921167294151266,
292
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
293
+ "eval_dim_768_cosine_precision@10": 0.4476190476190477,
294
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
295
+ "eval_dim_768_cosine_precision@5": 0.47619047619047616,
296
+ "eval_dim_768_cosine_recall@1": 0.08933150183150182,
297
+ "eval_dim_768_cosine_recall@10": 0.5401404151404151,
298
+ "eval_dim_768_cosine_recall@3": 0.24418498168498168,
299
+ "eval_dim_768_cosine_recall@5": 0.33951465201465203,
300
+ "eval_runtime": 6.9723,
301
+ "eval_samples_per_second": 0.0,
302
+ "eval_sequential_score": 0.51131642091388,
303
+ "eval_steps_per_second": 0.0,
304
+ "step": 14
305
+ }
306
+ ],
307
+ "logging_steps": 1,
308
+ "max_steps": 50,
309
+ "num_input_tokens_seen": 0,
310
+ "num_train_epochs": 10,
311
+ "save_steps": 500,
312
+ "stateful_callbacks": {
313
+ "EarlyStoppingCallback": {
314
+ "args": {
315
+ "early_stopping_patience": 3,
316
+ "early_stopping_threshold": 0.0
317
+ },
318
+ "attributes": {
319
+ "early_stopping_patience_counter": 0
320
+ }
321
+ },
322
+ "TrainerControl": {
323
+ "args": {
324
+ "should_epoch_stop": false,
325
+ "should_evaluate": false,
326
+ "should_log": false,
327
+ "should_save": true,
328
+ "should_training_stop": false
329
+ },
330
+ "attributes": {}
331
+ }
332
+ },
333
+ "total_flos": 0.0,
334
+ "train_batch_size": 8,
335
+ "trial_name": null,
336
+ "trial_params": null
337
+ }
checkpoint-14/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14debc6c3f8c5edee5db8d97a3a78a007d313a13e4b96f43026da543b59bef8c
3
+ size 6097
checkpoint-28/1_Pooling/config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "word_embedding_dimension": 1024,
3
+ "pooling_mode_cls_token": false,
4
+ "pooling_mode_mean_tokens": true,
5
+ "pooling_mode_max_tokens": false,
6
+ "pooling_mode_mean_sqrt_len_tokens": false,
7
+ "pooling_mode_weightedmean_tokens": false,
8
+ "pooling_mode_lasttoken": false,
9
+ "include_prompt": true
10
+ }
checkpoint-28/README.md ADDED
@@ -0,0 +1,1651 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license: apache-2.0
5
+ tags:
6
+ - sentence-transformers
7
+ - sentence-similarity
8
+ - feature-extraction
9
+ - dense
10
+ - generated_from_trainer
11
+ - dataset_size:82
12
+ - loss:MatryoshkaLoss
13
+ - loss:MultipleNegativesRankingLoss
14
+ base_model: intfloat/multilingual-e5-large
15
+ widget:
16
+ - source_sentence: When did the victims give away credentials?
17
+ sentences:
18
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
19
+
20
+
21
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
22
+ benefit, causes damage to another’s property by persuading someone to act, omit,
23
+ or tolerate something through the knowing misrepresentation of false facts as
24
+ true, or through the unlawful concealment or suppression of true facts, shall
25
+ be punished by imprisonment of at least three months, and if the damage caused
26
+ is particularly large, by imprisonment of at least two years."
27
+
28
+
29
+ From this provision it follows that, for the crime of fraud to be established,
30
+ the following elements are required:
31
+
32
+
33
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
34
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
35
+
36
+
37
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
38
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
39
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
40
+ to themselves or another; and
41
+
42
+
43
+ c) Damage to another person’s property, as defined under civil law, which must
44
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
45
+ not required that the person deceived and the person who suffered the damage be
46
+ the same individual.
47
+
48
+
49
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
50
+ relating to the past or present, and not to those that will occur in the future,
51
+ such as mere promises or contractual obligations. However, when such promises
52
+ or obligations are accompanied by false assurances and representations of other
53
+ false facts referring to the present or the past, in such a manner as to create
54
+ the impression of future fulfillment based on a false present situation fabricated
55
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
56
+ the crime of fraud is established.
57
+
58
+
59
+ The term “property” refers to the totality of a person’s economic assets that
60
+ possess monetary value, while damage to property means its reduction—specifically,
61
+ the difference between the monetary value the property had before the disposition
62
+ caused by the fraudulent conduct and the value remaining after it. Property damage
63
+ exists even if the victim possesses an active claim for restitution.
64
+
65
+
66
+ The time of commission of the fraud is considered to be the moment when the perpetrator
67
+ acted and completed their fraudulent conduct, namely when they made the false
68
+ representations that deceived the victim or a third party. Any subsequent moment
69
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
70
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
71
+ - 'Voice phishing involves manipulating victims over the phone. Attackers pose as
72
+ bank officials or authorities and use intimidation to extract financial details.
73
+
74
+
75
+ Scenario:
76
+
77
+ - Victims are coerced into giving away PINs, passwords, or other credentials under
78
+ false pretenses of legal or financial emergencies.'
79
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
80
+
81
+
82
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
83
+ benefit, causes damage to another’s property by persuading someone to act, omit,
84
+ or tolerate something through the knowing misrepresentation of false facts as
85
+ true, or through the unlawful concealment or suppression of true facts, shall
86
+ be punished by imprisonment of at least three months, and if the damage caused
87
+ is particularly large, by imprisonment of at least two years."
88
+
89
+
90
+ From this provision, it follows that, for the crime of fraud to be established,
91
+ the following elements are required:
92
+
93
+
94
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
95
+ pecuniary benefit, without requiring that the benefit actually materialize;
96
+
97
+
98
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
99
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
100
+ is deceived and performs an act, omission, or acquiescence; and
101
+
102
+
103
+ c) Damage to another’s property, according to civil law, which must be causally
104
+ connected to the perpetrator’s deceptive acts or omissions. It is not required
105
+ that the deceived person and the person who suffered the loss be the same.
106
+
107
+
108
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
109
+ relating to the past or present, and not to those that will occur in the future,
110
+ such as mere promises or contractual obligations. However, when such promises
111
+ or obligations are accompanied by false assurances and representations of other
112
+ false facts relating to the present or the past, in such a way as to create the
113
+ impression of future fulfillment, based on a false present situation fabricated
114
+ by the perpetrator—who has already made the decision not to fulfill their obligation—then
115
+ the crime of fraud is established.
116
+
117
+
118
+ The term “property” denotes the totality of a person’s economic assets possessing
119
+ monetary value, while damage to property refers to its reduction—specifically,
120
+ the difference between the property’s monetary value before the disposition caused
121
+ by the fraudulent conduct and its value afterward. Property damage exists even
122
+ if the victim has an active claim for its restitution.
123
+
124
+
125
+ The time of commission of fraud is considered to be the moment when the perpetrator
126
+ acted and completed the deceptive conduct, that is, when they made the false representations
127
+ which deceived the victim or a third party. Any later time at which the victim’s
128
+ financial loss occurred—thus completing the fraud—or the time when the harmful
129
+ act or omission of the deceived person took place, is irrelevant.
130
+
131
+
132
+ The reference to multiple modes of commission of fraud (i.e., both the misrepresentation
133
+ of false facts and the concealment of true ones) may create ambiguity and contradiction,
134
+ unless it is made clear from the overall findings that the offense was committed
135
+ in one particular manner, and that the reference to the other merely serves to
136
+ define the intent (mens rea) of the perpetrator—specifically, that the representations
137
+ were false.
138
+
139
+
140
+ Furthermore, a conviction must contain the specific and well-reasoned justification
141
+ required by Articles 93 paragraph 3 of the Constitution and 139 of the Code of
142
+ Criminal Procedure. The absence of such reasoning constitutes grounds for cassation
143
+ (appeal) under Article 510 paragraph 1(d) of the Code of Criminal Procedure, when
144
+ the judgment does not set out, with clarity, completeness, and consistency, the
145
+ factual circumstances established by the evidence, upon which the court based
146
+ its findings regarding the objective and subjective elements of the offense, the
147
+ evidence supporting those findings, and the legal reasoning through which those
148
+ facts were subsumed under the applicable substantive criminal provision.
149
+
150
+
151
+ For the existence of such reasoning, the explanatory and operative parts of the
152
+ decision may complement each other, as they form a single, unified whole.
153
+
154
+
155
+ The existence of intent (dolus) does not generally need to be specially justified,
156
+ since it is inherent in the will to bring about the factual circumstances constituting
157
+ the objective elements of the offense, and it is presumed from their realization
158
+ in each particular case—unless the law requires additional elements for criminal
159
+ liability, such as the act being committed with knowledge of a specific circumstance
160
+ (direct intent) or with the pursuit of a further purpose, i.e., the achievement
161
+ of an additional result (offenses requiring a special subjective element).
162
+
163
+
164
+ Furthermore, under Article 510 paragraph 1(e) of the Code of Criminal Procedure,
165
+ a misapplication of substantive criminal law also constitutes grounds for cassation.
166
+ Such misapplication occurs when the trial court incorrectly applies the law to
167
+ the facts it has found to be true, or when the violation occurs indirectly, namely
168
+ when the reasoning of the judgment—comprising the combination of its factual and
169
+ operative parts and relating to the elements and identity of the offense—contains
170
+ ambiguities, contradictions, or logical gaps, rendering it impossible to verify,
171
+ on appeal, whether the law was applied correctly. In such cases, the judgment
172
+ lacks a lawful basis.'
173
+ - source_sentence: What must be the outcome of the deception in relation to property
174
+ damage?
175
+ sentences:
176
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
177
+
178
+
179
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
180
+ benefit, causes damage to another’s property by persuading someone to act, omit,
181
+ or tolerate something through the knowing misrepresentation of false facts as
182
+ true, or through the unlawful concealment or suppression of true facts, shall
183
+ be punished by imprisonment of at least three months, and if the damage caused
184
+ is particularly large, by imprisonment of at least two years."
185
+
186
+
187
+ From this provision, it follows that, for the crime of fraud to be established,
188
+ the following elements are required:
189
+
190
+
191
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
192
+ pecuniary benefit, without requiring that the benefit actually materialize;
193
+
194
+
195
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
196
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
197
+ is deceived and performs an act, omission, or acquiescence; and
198
+
199
+
200
+ c) Damage to another’s property, according to civil law, which must be causally
201
+ connected to the perpetrator’s deceptive acts or omissions. It is not required
202
+ that the deceived person and the person who suffered the loss be the same.
203
+
204
+
205
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
206
+ relating to the past or present, and not to those that will occur in the future,
207
+ such as mere promises or contractual obligations. However, when such promises
208
+ or obligations are accompanied by false assurances and representations of other
209
+ false facts relating to the present or the past, in such a way as to create the
210
+ impression of future fulfillment, based on a false present situation fabricated
211
+ by the perpetrator—who has already made the decision not to fulfill their obligation—then
212
+ the crime of fraud is established.
213
+
214
+
215
+ The term “property” denotes the totality of a person’s economic assets possessing
216
+ monetary value, while damage to property refers to its reduction—specifically,
217
+ the difference between the property’s monetary value before the disposition caused
218
+ by the fraudulent conduct and its value afterward. Property damage exists even
219
+ if the victim has an active claim for its restitution.
220
+
221
+
222
+ The time of commission of fraud is considered to be the moment when the perpetrator
223
+ acted and completed the deceptive conduct, that is, when they made the false representations
224
+ which deceived the victim or a third party. Any later time at which the victim’s
225
+ financial loss occurred—thus completing the fraud—or the time when the harmful
226
+ act or omission of the deceived person took place, is irrelevant.
227
+
228
+
229
+ The reference to multiple modes of commission of fraud (i.e., both the misrepresentation
230
+ of false facts and the concealment of true ones) may create ambiguity and contradiction,
231
+ unless it is made clear from the overall findings that the offense was committed
232
+ in one particular manner, and that the reference to the other merely serves to
233
+ define the intent (mens rea) of the perpetrator—specifically, that the representations
234
+ were false.
235
+
236
+
237
+ Furthermore, a conviction must contain the specific and well-reasoned justification
238
+ required by Articles 93 paragraph 3 of the Constitution and 139 of the Code of
239
+ Criminal Procedure. The absence of such reasoning constitutes grounds for cassation
240
+ (appeal) under Article 510 paragraph 1(d) of the Code of Criminal Procedure, when
241
+ the judgment does not set out, with clarity, completeness, and consistency, the
242
+ factual circumstances established by the evidence, upon which the court based
243
+ its findings regarding the objective and subjective elements of the offense, the
244
+ evidence supporting those findings, and the legal reasoning through which those
245
+ facts were subsumed under the applicable substantive criminal provision.
246
+
247
+
248
+ For the existence of such reasoning, the explanatory and operative parts of the
249
+ decision may complement each other, as they form a single, unified whole.
250
+
251
+
252
+ The existence of intent (dolus) does not generally need to be specially justified,
253
+ since it is inherent in the will to bring about the factual circumstances constituting
254
+ the objective elements of the offense, and it is presumed from their realization
255
+ in each particular case—unless the law requires additional elements for criminal
256
+ liability, such as the act being committed with knowledge of a specific circumstance
257
+ (direct intent) or with the pursuit of a further purpose, i.e., the achievement
258
+ of an additional result (offenses requiring a special subjective element).
259
+
260
+
261
+ Furthermore, under Article 510 paragraph 1(e) of the Code of Criminal Procedure,
262
+ a misapplication of substantive criminal law also constitutes grounds for cassation.
263
+ Such misapplication occurs when the trial court incorrectly applies the law to
264
+ the facts it has found to be true, or when the violation occurs indirectly, namely
265
+ when the reasoning of the judgment—comprising the combination of its factual and
266
+ operative parts and relating to the elements and identity of the offense—contains
267
+ ambiguities, contradictions, or logical gaps, rendering it impossible to verify,
268
+ on appeal, whether the law was applied correctly. In such cases, the judgment
269
+ lacks a lawful basis.'
270
+ - 'According to Article 386 paragraph 1 of the Greek Penal Code,
271
+
272
+
273
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
274
+ benefit, causes damage to another’s property by persuading someone to act, omit,
275
+ or tolerate something through the knowing misrepresentation of false facts as
276
+ true, or through the unlawful concealment or suppression of true facts, shall
277
+ be punished by imprisonment of at least three months, and if the damage caused
278
+ is particularly large, by imprisonment of at least two years."
279
+
280
+
281
+ From these provisions, it follows that, for the crime of fraud to be established,
282
+ the following elements are required:
283
+
284
+
285
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
286
+ pecuniary benefit;
287
+
288
+
289
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
290
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
291
+ is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves
292
+ or another; and
293
+
294
+
295
+ c) Damage to another’s property, as defined under civil law, which must be causally
296
+ connected to the perpetrator’s deceptive acts.
297
+
298
+
299
+ From the above provisions, it is deduced that the crime of fraud is established
300
+ both objectively and subjectively through the knowing misrepresentation of false
301
+ facts as true, or the unlawful concealment or suppression of true ones, by which
302
+ another person is deceived and, as a result, performs an act, omission, or acquiescence
303
+ involving a disposition of property that directly and necessarily causes financial
304
+ damage to the deceived person or another, with the intent that the perpetrator
305
+ or another gain an unlawful benefit. It is irrelevant whether this intended benefit
306
+ was ultimately achieved.
307
+
308
+
309
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
310
+ relating to the past or present, and not to those expected to occur in the future,
311
+ such as mere promises or contractual obligations. The false fact must have existed
312
+ in the past or must be a present circumstance at the time it is asserted, and
313
+ cannot relate to the future.
314
+
315
+
316
+ However, when future circumstances—that is, promises or contractual obligations—are
317
+ accompanied by false assurances and representations of other false facts referring
318
+ to the present or past, in such a way as to create the impression of future fulfillment,
319
+ based on a false present situation or supposed ability of the perpetrator, who
320
+ had already made the decision not to fulfill their obligation, then the crime
321
+ of fraud is established.'
322
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
323
+
324
+
325
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
326
+ benefit, causes damage to another person’s property by persuading someone to act,
327
+ omit, or tolerate something through the knowing misrepresentation of false facts
328
+ as true, or through the unlawful concealment or suppression of true facts, shall
329
+ be punished by imprisonment of at least three months, and if the damage caused
330
+ is particularly large, by imprisonment of at least two years."
331
+
332
+
333
+ From this provision, it follows that for the crime of fraud to be established,
334
+ the following elements are required:
335
+
336
+
337
+ a) Intent of the perpetrator to obtain for themselves or another an unlawful pecuniary
338
+ benefit, regardless of whether this benefit was actually realized;
339
+
340
+
341
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
342
+ or suppression of true facts, as a result of which, as a causal factor, someone
343
+ is deceived and acts in a way that is detrimental to themselves or another (by
344
+ an act, omission, or acquiescence); and
345
+
346
+
347
+ c) Damage to another’s property, in the sense recognized by civil law, which must
348
+ be causally linked to the fraudulent conduct (the deceptive act or omission of
349
+ the perpetrator) and to the resulting deception of the person who made the property
350
+ disposition. It is not required that the person deceived be the same person who
351
+ suffered the damage.
352
+
353
+
354
+ Property damage exists when there is a reduction or deterioration in the victim’s
355
+ assets, even if the victim has an active claim to restitution. However, as an
356
+ element of the objective aspect of the crime of fraud, the damage must be the
357
+ direct, necessary, and exclusive result of the property disposition—namely, the
358
+ act, omission, or acquiescence performed by the person deceived by the perpetrator’s
359
+ fraudulent conduct.
360
+
361
+
362
+ There must therefore be a causal connection between the perpetrator’s deceptive
363
+ behavior and the deception it caused, as well as between this deception and the
364
+ resulting property damage, which must be the direct, necessary, and exclusive
365
+ outcome of the deception and of the act, omission, or acquiescence of the deceived
366
+ person.
367
+
368
+
369
+ The term “facts” refers to real circumstances relating to the past or present,
370
+ and not to those expected to occur in the future, such as mere promises or contractual
371
+ obligations. However, when such promises or obligations are accompanied by false
372
+ assurances and representations of other false facts relating to the present or
373
+ the past, in such a way as to create the impression of future fulfillment, based
374
+ on the false present situation presented by a perpetrator who has already made
375
+ the decision not to fulfill their obligation, then the crime of fraud is established.
376
+
377
+
378
+ The time of commission of the fraud is considered to be the moment when the perpetrator
379
+ acted and completed their deceptive conduct—that is, when they made the false
380
+ representations that deceived the victim or a third party. Any later time at which
381
+ the victim’s financial loss actually occurred—thus completing the fraud—or the
382
+ time when the deceived person performed the harmful act or omission, is irrelevant.'
383
+ - source_sentence: How are victims tricked in email phishing scams?
384
+ sentences:
385
+ - 'According to Article 386 paragraph 1 of the Greek Penal Code,
386
+
387
+
388
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
389
+ benefit, causes damage to another’s property by persuading someone to act, omit,
390
+ or tolerate something through the knowing misrepresentation of false facts as
391
+ true, or through the unlawful concealment or suppression of true facts, shall
392
+ be punished by imprisonment of at least three months, and if the damage caused
393
+ is particularly large, by imprisonment of at least two years."
394
+
395
+
396
+ From these provisions, it follows that, for the crime of fraud to be established,
397
+ the following elements are required:
398
+
399
+
400
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
401
+ pecuniary benefit;
402
+
403
+
404
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
405
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
406
+ is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves
407
+ or another; and
408
+
409
+
410
+ c) Damage to another’s property, as defined under civil law, which must be causally
411
+ connected to the perpetrator’s deceptive acts.
412
+
413
+
414
+ From the above provisions, it is deduced that the crime of fraud is established
415
+ both objectively and subjectively through the knowing misrepresentation of false
416
+ facts as true, or the unlawful concealment or suppression of true ones, by which
417
+ another person is deceived and, as a result, performs an act, omission, or acquiescence
418
+ involving a disposition of property that directly and necessarily causes financial
419
+ damage to the deceived person or another, with the intent that the perpetrator
420
+ or another gain an unlawful benefit. It is irrelevant whether this intended benefit
421
+ was ultimately achieved.
422
+
423
+
424
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
425
+ relating to the past or present, and not to those expected to occur in the future,
426
+ such as mere promises or contractual obligations. The false fact must have existed
427
+ in the past or must be a present circumstance at the time it is asserted, and
428
+ cannot relate to the future.
429
+
430
+
431
+ However, when future circumstances—that is, promises or contractual obligations—are
432
+ accompanied by false assurances and representations of other false facts referring
433
+ to the present or past, in such a way as to create the impression of future fulfillment,
434
+ based on a false present situation or supposed ability of the perpetrator, who
435
+ had already made the decision not to fulfill their obligation, then the crime
436
+ of fraud is established.'
437
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
438
+
439
+
440
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
441
+ benefit, causes damage to another’s property by persuading someone to act, omit,
442
+ or tolerate something through the knowing misrepresentation of false facts as
443
+ true, or through the unlawful concealment or suppression of true facts, shall
444
+ be punished by imprisonment of at least three months, and if the damage caused
445
+ is particularly large, by imprisonment of at least two years."
446
+
447
+
448
+ From this provision it follows that, for the crime of fraud to be established,
449
+ the following elements are required:
450
+
451
+
452
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
453
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
454
+
455
+
456
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
457
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
458
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
459
+ to themselves or another; and
460
+
461
+
462
+ c) Damage to another person’s property, as defined under civil law, which must
463
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
464
+ not required that the person deceived and the person who suffered the damage be
465
+ the same individual.
466
+
467
+
468
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
469
+ relating to the past or present, and not to those that will occur in the future,
470
+ such as mere promises or contractual obligations. However, when such promises
471
+ or obligations are accompanied by false assurances and representations of other
472
+ false facts referring to the present or the past, in such a manner as to create
473
+ the impression of future fulfillment based on a false present situation fabricated
474
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
475
+ the crime of fraud is established.
476
+
477
+
478
+ The term “property” refers to the totality of a person’s economic assets that
479
+ possess monetary value, while damage to property means its reduction—specifically,
480
+ the difference between the monetary value the property had before the disposition
481
+ caused by the fraudulent conduct and the value remaining after it. Property damage
482
+ exists even if the victim possesses an active claim for restitution.
483
+
484
+
485
+ The time of commission of the fraud is considered to be the moment when the perpetrator
486
+ acted and completed their fraudulent conduct, namely when they made the false
487
+ representations that deceived the victim or a third party. Any subsequent moment
488
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
489
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
490
+ - 'Email phishing is a type of identity theft scam conducted via email or SMS. The
491
+ attacker uses social engineering tactics such as impersonating trusted entities
492
+ and inducing urgency. Victims are tricked into disclosing personal information
493
+ or downloading malware.
494
+
495
+
496
+ Scenarios:
497
+
498
+ - Scenario 1: Emails impersonating high-ranking executives accuse victims of crimes
499
+ to coerce them into revealing information or opening malware-laden attachments.
500
+
501
+ - Scenario 2: Emails/SMS from fake banks or authorities alert victims of data
502
+ breaches, directing them to spoofed websites to input credentials.
503
+
504
+ - Scenario 3: SMS messages deliver disguised malware apps that harvest sensitive
505
+ data.
506
+
507
+ - Scenario 4: SMS links lead to pharming sites that mimic trusted brands and steal
508
+ login data through fake pop-ups.'
509
+ - source_sentence: What circumstances do the term 'facts' refer to within the meaning
510
+ of the provision?
511
+ sentences:
512
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
513
+
514
+
515
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
516
+ benefit, causes damage to another person’s property by persuading someone to act,
517
+ omit, or tolerate something through the knowing misrepresentation of false facts
518
+ as true, or through the unlawful concealment or suppression of true facts, shall
519
+ be punished by imprisonment of at least three months, and if the damage caused
520
+ is particularly large, by imprisonment of at least two years."
521
+
522
+
523
+ From this provision, it follows that for the crime of fraud to be established,
524
+ the following elements are required:
525
+
526
+
527
+ a) Intent of the perpetrator to obtain for themselves or another an unlawful pecuniary
528
+ benefit, regardless of whether this benefit was actually realized;
529
+
530
+
531
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
532
+ or suppression of true facts, as a result of which, as a causal factor, someone
533
+ is deceived and acts in a way that is detrimental to themselves or another (by
534
+ an act, omission, or acquiescence); and
535
+
536
+
537
+ c) Damage to another’s property, in the sense recognized by civil law, which must
538
+ be causally linked to the fraudulent conduct (the deceptive act or omission of
539
+ the perpetrator) and to the resulting deception of the person who made the property
540
+ disposition. It is not required that the person deceived be the same person who
541
+ suffered the damage.
542
+
543
+
544
+ Property damage exists when there is a reduction or deterioration in the victim’s
545
+ assets, even if the victim has an active claim to restitution. However, as an
546
+ element of the objective aspect of the crime of fraud, the damage must be the
547
+ direct, necessary, and exclusive result of the property disposition—namely, the
548
+ act, omission, or acquiescence performed by the person deceived by the perpetrator’s
549
+ fraudulent conduct.
550
+
551
+
552
+ There must therefore be a causal connection between the perpetrator’s deceptive
553
+ behavior and the deception it caused, as well as between this deception and the
554
+ resulting property damage, which must be the direct, necessary, and exclusive
555
+ outcome of the deception and of the act, omission, or acquiescence of the deceived
556
+ person.
557
+
558
+
559
+ The term “facts” refers to real circumstances relating to the past or present,
560
+ and not to those expected to occur in the future, such as mere promises or contractual
561
+ obligations. However, when such promises or obligations are accompanied by false
562
+ assurances and representations of other false facts relating to the present or
563
+ the past, in such a way as to create the impression of future fulfillment, based
564
+ on the false present situation presented by a perpetrator who has already made
565
+ the decision not to fulfill their obligation, then the crime of fraud is established.
566
+
567
+
568
+ The time of commission of the fraud is considered to be the moment when the perpetrator
569
+ acted and completed their deceptive conduct—that is, when they made the false
570
+ representations that deceived the victim or a third party. Any later time at which
571
+ the victim’s financial loss actually occurred—thus completing the fraud—or the
572
+ time when the deceived person performed the harmful act or omission, is irrelevant.'
573
+ - '1. Anyone who, by knowingly presenting false facts as true or by unlawfully concealing
574
+ or withholding true facts, damages another person''s property by persuading someone
575
+ to act, omission, or tolerance with the aim of obtaining, for themselves or another,
576
+ an unlawful financial gain from the damage to that property shall be punished
577
+ with imprisonment, "and if the damage caused is particularly great, with imprisonment
578
+ of at least three (3) months and a fine." .
579
+
580
+ If the damage caused exceeds a total of one hundred and twenty thousand (120,000)
581
+ euros, imprisonment of up to ten (10) years and a fine shall be imposed.
582
+
583
+ 2. If the fraud is directed directly against the legal entity of the Greek State,
584
+ legal entities governed by public law, or local government organizations, and
585
+ the damage caused exceeds a total of one hundred and twenty thousand (120,000)
586
+ euros, a prison sentence of at least ten (10) years and a fine of up to one thousand
587
+ (1,000) daily units shall be imposed. This offense shall be time-barred after
588
+ twenty (20) years.
589
+
590
+ '
591
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
592
+
593
+
594
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
595
+ benefit, causes damage to another’s property by persuading someone to act, omit,
596
+ or tolerate something through the knowing misrepresentation of false facts as
597
+ true, or through the unlawful concealment or suppression of true facts, shall
598
+ be punished by imprisonment of at least three months, and if the damage caused
599
+ is particularly large, by imprisonment of at least two years."
600
+
601
+
602
+ From this provision it follows that, for the crime of fraud to be established,
603
+ the following elements are required:
604
+
605
+
606
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
607
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
608
+
609
+
610
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
611
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
612
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
613
+ to themselves or another; and
614
+
615
+
616
+ c) Damage to another person’s property, as defined under civil law, which must
617
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
618
+ not required that the person deceived and the person who suffered the damage be
619
+ the same individual.
620
+
621
+
622
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
623
+ relating to the past or present, and not to those that will occur in the future,
624
+ such as mere promises or contractual obligations. However, when such promises
625
+ or obligations are accompanied by false assurances and representations of other
626
+ false facts referring to the present or the past, in such a manner as to create
627
+ the impression of future fulfillment based on a false present situation fabricated
628
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
629
+ the crime of fraud is established.
630
+
631
+
632
+ The term “property” refers to the totality of a person’s economic assets that
633
+ possess monetary value, while damage to property means its reduction—specifically,
634
+ the difference between the monetary value the property had before the disposition
635
+ caused by the fraudulent conduct and the value remaining after it. Property damage
636
+ exists even if the victim possesses an active claim for restitution.
637
+
638
+
639
+ The time of commission of the fraud is considered to be the moment when the perpetrator
640
+ acted and completed their fraudulent conduct, namely when they made the false
641
+ representations that deceived the victim or a third party. Any subsequent moment
642
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
643
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
644
+ - source_sentence: When is the time of commission of the fraud considered?
645
+ sentences:
646
+ - 'Spear phishing targets specific individuals or employees within an organization
647
+ using personalized, deceptive emails. Unlike mass phishing, these emails are crafted
648
+ to seem familiar and urgent.
649
+
650
+
651
+ Scenarios:
652
+
653
+ - CEO Fraud: Attackers impersonate executives to extract financial or sensitive
654
+ data from employees.
655
+
656
+ - Whaling: High-ranking executives are targeted using tailored fraud emails that
657
+ press for immediate action without verification.'
658
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
659
+
660
+
661
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
662
+ benefit, causes damage to another’s property by persuading someone to act, omit,
663
+ or tolerate something through the knowing misrepresentation of false facts as
664
+ true, or through the unlawful concealment or suppression of true facts, shall
665
+ be punished by imprisonment of at least three months, and if the damage caused
666
+ is particularly large, by imprisonment of at least two years."
667
+
668
+
669
+ From this provision it follows that, for the crime of fraud to be established,
670
+ the following elements are required:
671
+
672
+
673
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
674
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
675
+
676
+
677
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
678
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
679
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
680
+ to themselves or another; and
681
+
682
+
683
+ c) Damage to another person’s property, as defined under civil law, which must
684
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
685
+ not required that the person deceived and the person who suffered the damage be
686
+ the same individual.
687
+
688
+
689
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
690
+ relating to the past or present, and not to those that will occur in the future,
691
+ such as mere promises or contractual obligations. However, when such promises
692
+ or obligations are accompanied by false assurances and representations of other
693
+ false facts referring to the present or the past, in such a manner as to create
694
+ the impression of future fulfillment based on a false present situation fabricated
695
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
696
+ the crime of fraud is established.
697
+
698
+
699
+ The term “property” refers to the totality of a person’s economic assets that
700
+ possess monetary value, while damage to property means its reduction—specifically,
701
+ the difference between the monetary value the property had before the disposition
702
+ caused by the fraudulent conduct and the value remaining after it. Property damage
703
+ exists even if the victim possesses an active claim for restitution.
704
+
705
+
706
+ The time of commission of the fraud is considered to be the moment when the perpetrator
707
+ acted and completed their fraudulent conduct, namely when they made the false
708
+ representations that deceived the victim or a third party. Any subsequent moment
709
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
710
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
711
+ - 'According to Article 386 paragraph 1 of the Greek Penal Code,
712
+
713
+
714
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
715
+ benefit, causes damage to another’s property by persuading someone to act, omit,
716
+ or tolerate something through the knowing misrepresentation of false facts as
717
+ true, or through the unlawful concealment or suppression of true facts, shall
718
+ be punished by imprisonment of at least three months, and if the damage caused
719
+ is particularly large, by imprisonment of at least two years."
720
+
721
+
722
+ From these provisions, it follows that, for the crime of fraud to be established,
723
+ the following elements are required:
724
+
725
+
726
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
727
+ pecuniary benefit;
728
+
729
+
730
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
731
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
732
+ is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves
733
+ or another; and
734
+
735
+
736
+ c) Damage to another’s property, as defined under civil law, which must be causally
737
+ connected to the perpetrator’s deceptive acts.
738
+
739
+
740
+ From the above provisions, it is deduced that the crime of fraud is established
741
+ both objectively and subjectively through the knowing misrepresentation of false
742
+ facts as true, or the unlawful concealment or suppression of true ones, by which
743
+ another person is deceived and, as a result, performs an act, omission, or acquiescence
744
+ involving a disposition of property that directly and necessarily causes financial
745
+ damage to the deceived person or another, with the intent that the perpetrator
746
+ or another gain an unlawful benefit. It is irrelevant whether this intended benefit
747
+ was ultimately achieved.
748
+
749
+
750
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
751
+ relating to the past or present, and not to those expected to occur in the future,
752
+ such as mere promises or contractual obligations. The false fact must have existed
753
+ in the past or must be a present circumstance at the time it is asserted, and
754
+ cannot relate to the future.
755
+
756
+
757
+ However, when future circumstances—that is, promises or contractual obligations—are
758
+ accompanied by false assurances and representations of other false facts referring
759
+ to the present or past, in such a way as to create the impression of future fulfillment,
760
+ based on a false present situation or supposed ability of the perpetrator, who
761
+ had already made the decision not to fulfill their obligation, then the crime
762
+ of fraud is established.'
763
+ pipeline_tag: sentence-similarity
764
+ library_name: sentence-transformers
765
+ metrics:
766
+ - cosine_accuracy@1
767
+ - cosine_accuracy@3
768
+ - cosine_accuracy@5
769
+ - cosine_accuracy@10
770
+ - cosine_precision@1
771
+ - cosine_precision@3
772
+ - cosine_precision@5
773
+ - cosine_precision@10
774
+ - cosine_recall@1
775
+ - cosine_recall@3
776
+ - cosine_recall@5
777
+ - cosine_recall@10
778
+ - cosine_ndcg@10
779
+ - cosine_mrr@10
780
+ - cosine_map@100
781
+ model-index:
782
+ - name: multilingual_e5_large Finetuned on Data
783
+ results:
784
+ - task:
785
+ type: information-retrieval
786
+ name: Information Retrieval
787
+ dataset:
788
+ name: dim 1024
789
+ type: dim_1024
790
+ metrics:
791
+ - type: cosine_accuracy@1
792
+ value: 0.47619047619047616
793
+ name: Cosine Accuracy@1
794
+ - type: cosine_accuracy@3
795
+ value: 0.47619047619047616
796
+ name: Cosine Accuracy@3
797
+ - type: cosine_accuracy@5
798
+ value: 0.47619047619047616
799
+ name: Cosine Accuracy@5
800
+ - type: cosine_accuracy@10
801
+ value: 0.5714285714285714
802
+ name: Cosine Accuracy@10
803
+ - type: cosine_precision@1
804
+ value: 0.47619047619047616
805
+ name: Cosine Precision@1
806
+ - type: cosine_precision@3
807
+ value: 0.4603174603174603
808
+ name: Cosine Precision@3
809
+ - type: cosine_precision@5
810
+ value: 0.419047619047619
811
+ name: Cosine Precision@5
812
+ - type: cosine_precision@10
813
+ value: 0.4
814
+ name: Cosine Precision@10
815
+ - type: cosine_recall@1
816
+ value: 0.07822039072039072
817
+ name: Cosine Recall@1
818
+ - type: cosine_recall@3
819
+ value: 0.21085164835164832
820
+ name: Cosine Recall@3
821
+ - type: cosine_recall@5
822
+ value: 0.27602258852258854
823
+ name: Cosine Recall@5
824
+ - type: cosine_recall@10
825
+ value: 0.4449023199023199
826
+ name: Cosine Recall@10
827
+ - type: cosine_ndcg@10
828
+ value: 0.5159384546892658
829
+ name: Cosine Ndcg@10
830
+ - type: cosine_mrr@10
831
+ value: 0.49092970521541945
832
+ name: Cosine Mrr@10
833
+ - type: cosine_map@100
834
+ value: 0.6149109740313521
835
+ name: Cosine Map@100
836
+ - task:
837
+ type: information-retrieval
838
+ name: Information Retrieval
839
+ dataset:
840
+ name: dim 768
841
+ type: dim_768
842
+ metrics:
843
+ - type: cosine_accuracy@1
844
+ value: 0.5238095238095238
845
+ name: Cosine Accuracy@1
846
+ - type: cosine_accuracy@3
847
+ value: 0.5238095238095238
848
+ name: Cosine Accuracy@3
849
+ - type: cosine_accuracy@5
850
+ value: 0.5238095238095238
851
+ name: Cosine Accuracy@5
852
+ - type: cosine_accuracy@10
853
+ value: 0.5714285714285714
854
+ name: Cosine Accuracy@10
855
+ - type: cosine_precision@1
856
+ value: 0.5238095238095238
857
+ name: Cosine Precision@1
858
+ - type: cosine_precision@3
859
+ value: 0.5079365079365079
860
+ name: Cosine Precision@3
861
+ - type: cosine_precision@5
862
+ value: 0.4666666666666666
863
+ name: Cosine Precision@5
864
+ - type: cosine_precision@10
865
+ value: 0.4238095238095238
866
+ name: Cosine Precision@10
867
+ - type: cosine_recall@1
868
+ value: 0.08218864468864469
869
+ name: Cosine Recall@1
870
+ - type: cosine_recall@3
871
+ value: 0.22275641025641024
872
+ name: Cosine Recall@3
873
+ - type: cosine_recall@5
874
+ value: 0.2958638583638584
875
+ name: Cosine Recall@5
876
+ - type: cosine_recall@10
877
+ value: 0.46474358974358976
878
+ name: Cosine Recall@10
879
+ - type: cosine_ndcg@10
880
+ value: 0.5468399582764966
881
+ name: Cosine Ndcg@10
882
+ - type: cosine_mrr@10
883
+ value: 0.5306122448979591
884
+ name: Cosine Mrr@10
885
+ - type: cosine_map@100
886
+ value: 0.6351788392177582
887
+ name: Cosine Map@100
888
+ - task:
889
+ type: information-retrieval
890
+ name: Information Retrieval
891
+ dataset:
892
+ name: dim 512
893
+ type: dim_512
894
+ metrics:
895
+ - type: cosine_accuracy@1
896
+ value: 0.47619047619047616
897
+ name: Cosine Accuracy@1
898
+ - type: cosine_accuracy@3
899
+ value: 0.47619047619047616
900
+ name: Cosine Accuracy@3
901
+ - type: cosine_accuracy@5
902
+ value: 0.47619047619047616
903
+ name: Cosine Accuracy@5
904
+ - type: cosine_accuracy@10
905
+ value: 0.5238095238095238
906
+ name: Cosine Accuracy@10
907
+ - type: cosine_precision@1
908
+ value: 0.47619047619047616
909
+ name: Cosine Precision@1
910
+ - type: cosine_precision@3
911
+ value: 0.4603174603174603
912
+ name: Cosine Precision@3
913
+ - type: cosine_precision@5
914
+ value: 0.419047619047619
915
+ name: Cosine Precision@5
916
+ - type: cosine_precision@10
917
+ value: 0.3761904761904762
918
+ name: Cosine Precision@10
919
+ - type: cosine_recall@1
920
+ value: 0.07822039072039072
921
+ name: Cosine Recall@1
922
+ - type: cosine_recall@3
923
+ value: 0.21085164835164832
924
+ name: Cosine Recall@3
925
+ - type: cosine_recall@5
926
+ value: 0.27602258852258854
927
+ name: Cosine Recall@5
928
+ - type: cosine_recall@10
929
+ value: 0.42506105006105005
930
+ name: Cosine Recall@10
931
+ - type: cosine_ndcg@10
932
+ value: 0.49922091065744895
933
+ name: Cosine Ndcg@10
934
+ - type: cosine_mrr@10
935
+ value: 0.48299319727891155
936
+ name: Cosine Mrr@10
937
+ - type: cosine_map@100
938
+ value: 0.5978106306698094
939
+ name: Cosine Map@100
940
+ - task:
941
+ type: information-retrieval
942
+ name: Information Retrieval
943
+ dataset:
944
+ name: dim 256
945
+ type: dim_256
946
+ metrics:
947
+ - type: cosine_accuracy@1
948
+ value: 0.5238095238095238
949
+ name: Cosine Accuracy@1
950
+ - type: cosine_accuracy@3
951
+ value: 0.5238095238095238
952
+ name: Cosine Accuracy@3
953
+ - type: cosine_accuracy@5
954
+ value: 0.5238095238095238
955
+ name: Cosine Accuracy@5
956
+ - type: cosine_accuracy@10
957
+ value: 0.5714285714285714
958
+ name: Cosine Accuracy@10
959
+ - type: cosine_precision@1
960
+ value: 0.5238095238095238
961
+ name: Cosine Precision@1
962
+ - type: cosine_precision@3
963
+ value: 0.5079365079365079
964
+ name: Cosine Precision@3
965
+ - type: cosine_precision@5
966
+ value: 0.4666666666666666
967
+ name: Cosine Precision@5
968
+ - type: cosine_precision@10
969
+ value: 0.4238095238095239
970
+ name: Cosine Precision@10
971
+ - type: cosine_recall@1
972
+ value: 0.08005189255189255
973
+ name: Cosine Recall@1
974
+ - type: cosine_recall@3
975
+ value: 0.21634615384615385
976
+ name: Cosine Recall@3
977
+ - type: cosine_recall@5
978
+ value: 0.28518009768009767
979
+ name: Cosine Recall@5
980
+ - type: cosine_recall@10
981
+ value: 0.4433760683760684
982
+ name: Cosine Recall@10
983
+ - type: cosine_ndcg@10
984
+ value: 0.5468399582764966
985
+ name: Cosine Ndcg@10
986
+ - type: cosine_mrr@10
987
+ value: 0.5306122448979591
988
+ name: Cosine Mrr@10
989
+ - type: cosine_map@100
990
+ value: 0.6411393184007045
991
+ name: Cosine Map@100
992
+ - task:
993
+ type: information-retrieval
994
+ name: Information Retrieval
995
+ dataset:
996
+ name: dim 128
997
+ type: dim_128
998
+ metrics:
999
+ - type: cosine_accuracy@1
1000
+ value: 0.47619047619047616
1001
+ name: Cosine Accuracy@1
1002
+ - type: cosine_accuracy@3
1003
+ value: 0.47619047619047616
1004
+ name: Cosine Accuracy@3
1005
+ - type: cosine_accuracy@5
1006
+ value: 0.47619047619047616
1007
+ name: Cosine Accuracy@5
1008
+ - type: cosine_accuracy@10
1009
+ value: 0.5238095238095238
1010
+ name: Cosine Accuracy@10
1011
+ - type: cosine_precision@1
1012
+ value: 0.47619047619047616
1013
+ name: Cosine Precision@1
1014
+ - type: cosine_precision@3
1015
+ value: 0.4603174603174603
1016
+ name: Cosine Precision@3
1017
+ - type: cosine_precision@5
1018
+ value: 0.419047619047619
1019
+ name: Cosine Precision@5
1020
+ - type: cosine_precision@10
1021
+ value: 0.3761904761904762
1022
+ name: Cosine Precision@10
1023
+ - type: cosine_recall@1
1024
+ value: 0.07822039072039072
1025
+ name: Cosine Recall@1
1026
+ - type: cosine_recall@3
1027
+ value: 0.21085164835164832
1028
+ name: Cosine Recall@3
1029
+ - type: cosine_recall@5
1030
+ value: 0.27602258852258854
1031
+ name: Cosine Recall@5
1032
+ - type: cosine_recall@10
1033
+ value: 0.42506105006105005
1034
+ name: Cosine Recall@10
1035
+ - type: cosine_ndcg@10
1036
+ value: 0.49922091065744895
1037
+ name: Cosine Ndcg@10
1038
+ - type: cosine_mrr@10
1039
+ value: 0.48299319727891155
1040
+ name: Cosine Mrr@10
1041
+ - type: cosine_map@100
1042
+ value: 0.6025310247157158
1043
+ name: Cosine Map@100
1044
+ - task:
1045
+ type: information-retrieval
1046
+ name: Information Retrieval
1047
+ dataset:
1048
+ name: dim 64
1049
+ type: dim_64
1050
+ metrics:
1051
+ - type: cosine_accuracy@1
1052
+ value: 0.47619047619047616
1053
+ name: Cosine Accuracy@1
1054
+ - type: cosine_accuracy@3
1055
+ value: 0.47619047619047616
1056
+ name: Cosine Accuracy@3
1057
+ - type: cosine_accuracy@5
1058
+ value: 0.47619047619047616
1059
+ name: Cosine Accuracy@5
1060
+ - type: cosine_accuracy@10
1061
+ value: 0.5238095238095238
1062
+ name: Cosine Accuracy@10
1063
+ - type: cosine_precision@1
1064
+ value: 0.47619047619047616
1065
+ name: Cosine Precision@1
1066
+ - type: cosine_precision@3
1067
+ value: 0.4603174603174603
1068
+ name: Cosine Precision@3
1069
+ - type: cosine_precision@5
1070
+ value: 0.419047619047619
1071
+ name: Cosine Precision@5
1072
+ - type: cosine_precision@10
1073
+ value: 0.3761904761904762
1074
+ name: Cosine Precision@10
1075
+ - type: cosine_recall@1
1076
+ value: 0.07822039072039072
1077
+ name: Cosine Recall@1
1078
+ - type: cosine_recall@3
1079
+ value: 0.21085164835164832
1080
+ name: Cosine Recall@3
1081
+ - type: cosine_recall@5
1082
+ value: 0.27602258852258854
1083
+ name: Cosine Recall@5
1084
+ - type: cosine_recall@10
1085
+ value: 0.42506105006105005
1086
+ name: Cosine Recall@10
1087
+ - type: cosine_ndcg@10
1088
+ value: 0.49922091065744895
1089
+ name: Cosine Ndcg@10
1090
+ - type: cosine_mrr@10
1091
+ value: 0.48299319727891155
1092
+ name: Cosine Mrr@10
1093
+ - type: cosine_map@100
1094
+ value: 0.5960251374266525
1095
+ name: Cosine Map@100
1096
+ ---
1097
+
1098
+ # multilingual_e5_large Finetuned on Data
1099
+
1100
+ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large). 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.
1101
+
1102
+ ## Model Details
1103
+
1104
+ ### Model Description
1105
+ - **Model Type:** Sentence Transformer
1106
+ - **Base model:** [intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large) <!-- at revision 0dc5580a448e4284468b8909bae50fa925907bc5 -->
1107
+ - **Maximum Sequence Length:** 512 tokens
1108
+ - **Output Dimensionality:** 1024 dimensions
1109
+ - **Similarity Function:** Cosine Similarity
1110
+ <!-- - **Training Dataset:** Unknown -->
1111
+ - **Language:** en
1112
+ - **License:** apache-2.0
1113
+
1114
+ ### Model Sources
1115
+
1116
+ - **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
1117
+ - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
1118
+ - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
1119
+
1120
+ ### Full Model Architecture
1121
+
1122
+ ```
1123
+ SentenceTransformer(
1124
+ (0): Transformer({'max_seq_length': 512, 'do_lower_case': False, 'architecture': 'XLMRobertaModel'})
1125
+ (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
1126
+ (2): Normalize()
1127
+ )
1128
+ ```
1129
+
1130
+ ## Usage
1131
+
1132
+ ### Direct Usage (Sentence Transformers)
1133
+
1134
+ First install the Sentence Transformers library:
1135
+
1136
+ ```bash
1137
+ pip install -U sentence-transformers
1138
+ ```
1139
+
1140
+ Then you can load this model and run inference.
1141
+ ```python
1142
+ from sentence_transformers import SentenceTransformer
1143
+
1144
+ # Download from the 🤗 Hub
1145
+ model = SentenceTransformer("sentence_transformers_model_id")
1146
+ # Run inference
1147
+ sentences = [
1148
+ 'When is the time of commission of the fraud considered?',
1149
+ 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom this provision it follows that, for the crime of fraud to be established, the following elements are required:\n\na) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, without it being necessary that the benefit actually materialize;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence that is detrimental to themselves or another; and\n\nc) Damage to another person’s property, as defined under civil law, which must be causally linked to the deceptive acts or omissions of the perpetrator. It is not required that the person deceived and the person who suffered the damage be the same individual.\n\nThe term “facts”, within the meaning of the above provision, refers to real circumstances relating to the past or present, and not to those that will occur in the future, such as mere promises or contractual obligations. However, when such promises or obligations are accompanied by false assurances and representations of other false facts referring to the present or the past, in such a manner as to create the impression of future fulfillment based on a false present situation fabricated by the perpetrator, who has already formed the decision not to fulfill their obligation, the crime of fraud is established.\n\nThe term “property” refers to the totality of a person’s economic assets that possess monetary value, while damage to property means its reduction—specifically, the difference between the monetary value the property had before the disposition caused by the fraudulent conduct and the value remaining after it. Property damage exists even if the victim possesses an active claim for restitution.\n\nThe time of commission of the fraud is considered to be the moment when the perpetrator acted and completed their fraudulent conduct, namely when they made the false representations that deceived the victim or a third party. Any subsequent moment at which the victim’s damage actually occurred—thereby completing the fraud—or the time when the victim carried out the harmful act or omission, is irrelevant.',
1150
+ 'Spear phishing targets specific individuals or employees within an organization using personalized, deceptive emails. Unlike mass phishing, these emails are crafted to seem familiar and urgent.\n\nScenarios:\n- CEO Fraud: Attackers impersonate executives to extract financial or sensitive data from employees.\n- Whaling: High-ranking executives are targeted using tailored fraud emails that press for immediate action without verification.',
1151
+ ]
1152
+ embeddings = model.encode(sentences)
1153
+ print(embeddings.shape)
1154
+ # [3, 1024]
1155
+
1156
+ # Get the similarity scores for the embeddings
1157
+ similarities = model.similarity(embeddings, embeddings)
1158
+ print(similarities)
1159
+ # tensor([[1.0000, 0.5637, 0.3101],
1160
+ # [0.5637, 1.0000, 0.3522],
1161
+ # [0.3101, 0.3522, 1.0000]])
1162
+ ```
1163
+
1164
+ <!--
1165
+ ### Direct Usage (Transformers)
1166
+
1167
+ <details><summary>Click to see the direct usage in Transformers</summary>
1168
+
1169
+ </details>
1170
+ -->
1171
+
1172
+ <!--
1173
+ ### Downstream Usage (Sentence Transformers)
1174
+
1175
+ You can finetune this model on your own dataset.
1176
+
1177
+ <details><summary>Click to expand</summary>
1178
+
1179
+ </details>
1180
+ -->
1181
+
1182
+ <!--
1183
+ ### Out-of-Scope Use
1184
+
1185
+ *List how the model may foreseeably be misused and address what users ought not to do with the model.*
1186
+ -->
1187
+
1188
+ ## Evaluation
1189
+
1190
+ ### Metrics
1191
+
1192
+ #### Information Retrieval
1193
+
1194
+ * Dataset: `dim_1024`
1195
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1196
+ ```json
1197
+ {
1198
+ "truncate_dim": 1024
1199
+ }
1200
+ ```
1201
+
1202
+ | Metric | Value |
1203
+ |:--------------------|:-----------|
1204
+ | cosine_accuracy@1 | 0.4762 |
1205
+ | cosine_accuracy@3 | 0.4762 |
1206
+ | cosine_accuracy@5 | 0.4762 |
1207
+ | cosine_accuracy@10 | 0.5714 |
1208
+ | cosine_precision@1 | 0.4762 |
1209
+ | cosine_precision@3 | 0.4603 |
1210
+ | cosine_precision@5 | 0.419 |
1211
+ | cosine_precision@10 | 0.4 |
1212
+ | cosine_recall@1 | 0.0782 |
1213
+ | cosine_recall@3 | 0.2109 |
1214
+ | cosine_recall@5 | 0.276 |
1215
+ | cosine_recall@10 | 0.4449 |
1216
+ | **cosine_ndcg@10** | **0.5159** |
1217
+ | cosine_mrr@10 | 0.4909 |
1218
+ | cosine_map@100 | 0.6149 |
1219
+
1220
+ #### Information Retrieval
1221
+
1222
+ * Dataset: `dim_768`
1223
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1224
+ ```json
1225
+ {
1226
+ "truncate_dim": 768
1227
+ }
1228
+ ```
1229
+
1230
+ | Metric | Value |
1231
+ |:--------------------|:-----------|
1232
+ | cosine_accuracy@1 | 0.5238 |
1233
+ | cosine_accuracy@3 | 0.5238 |
1234
+ | cosine_accuracy@5 | 0.5238 |
1235
+ | cosine_accuracy@10 | 0.5714 |
1236
+ | cosine_precision@1 | 0.5238 |
1237
+ | cosine_precision@3 | 0.5079 |
1238
+ | cosine_precision@5 | 0.4667 |
1239
+ | cosine_precision@10 | 0.4238 |
1240
+ | cosine_recall@1 | 0.0822 |
1241
+ | cosine_recall@3 | 0.2228 |
1242
+ | cosine_recall@5 | 0.2959 |
1243
+ | cosine_recall@10 | 0.4647 |
1244
+ | **cosine_ndcg@10** | **0.5468** |
1245
+ | cosine_mrr@10 | 0.5306 |
1246
+ | cosine_map@100 | 0.6352 |
1247
+
1248
+ #### Information Retrieval
1249
+
1250
+ * Dataset: `dim_512`
1251
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1252
+ ```json
1253
+ {
1254
+ "truncate_dim": 512
1255
+ }
1256
+ ```
1257
+
1258
+ | Metric | Value |
1259
+ |:--------------------|:-----------|
1260
+ | cosine_accuracy@1 | 0.4762 |
1261
+ | cosine_accuracy@3 | 0.4762 |
1262
+ | cosine_accuracy@5 | 0.4762 |
1263
+ | cosine_accuracy@10 | 0.5238 |
1264
+ | cosine_precision@1 | 0.4762 |
1265
+ | cosine_precision@3 | 0.4603 |
1266
+ | cosine_precision@5 | 0.419 |
1267
+ | cosine_precision@10 | 0.3762 |
1268
+ | cosine_recall@1 | 0.0782 |
1269
+ | cosine_recall@3 | 0.2109 |
1270
+ | cosine_recall@5 | 0.276 |
1271
+ | cosine_recall@10 | 0.4251 |
1272
+ | **cosine_ndcg@10** | **0.4992** |
1273
+ | cosine_mrr@10 | 0.483 |
1274
+ | cosine_map@100 | 0.5978 |
1275
+
1276
+ #### Information Retrieval
1277
+
1278
+ * Dataset: `dim_256`
1279
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1280
+ ```json
1281
+ {
1282
+ "truncate_dim": 256
1283
+ }
1284
+ ```
1285
+
1286
+ | Metric | Value |
1287
+ |:--------------------|:-----------|
1288
+ | cosine_accuracy@1 | 0.5238 |
1289
+ | cosine_accuracy@3 | 0.5238 |
1290
+ | cosine_accuracy@5 | 0.5238 |
1291
+ | cosine_accuracy@10 | 0.5714 |
1292
+ | cosine_precision@1 | 0.5238 |
1293
+ | cosine_precision@3 | 0.5079 |
1294
+ | cosine_precision@5 | 0.4667 |
1295
+ | cosine_precision@10 | 0.4238 |
1296
+ | cosine_recall@1 | 0.0801 |
1297
+ | cosine_recall@3 | 0.2163 |
1298
+ | cosine_recall@5 | 0.2852 |
1299
+ | cosine_recall@10 | 0.4434 |
1300
+ | **cosine_ndcg@10** | **0.5468** |
1301
+ | cosine_mrr@10 | 0.5306 |
1302
+ | cosine_map@100 | 0.6411 |
1303
+
1304
+ #### Information Retrieval
1305
+
1306
+ * Dataset: `dim_128`
1307
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1308
+ ```json
1309
+ {
1310
+ "truncate_dim": 128
1311
+ }
1312
+ ```
1313
+
1314
+ | Metric | Value |
1315
+ |:--------------------|:-----------|
1316
+ | cosine_accuracy@1 | 0.4762 |
1317
+ | cosine_accuracy@3 | 0.4762 |
1318
+ | cosine_accuracy@5 | 0.4762 |
1319
+ | cosine_accuracy@10 | 0.5238 |
1320
+ | cosine_precision@1 | 0.4762 |
1321
+ | cosine_precision@3 | 0.4603 |
1322
+ | cosine_precision@5 | 0.419 |
1323
+ | cosine_precision@10 | 0.3762 |
1324
+ | cosine_recall@1 | 0.0782 |
1325
+ | cosine_recall@3 | 0.2109 |
1326
+ | cosine_recall@5 | 0.276 |
1327
+ | cosine_recall@10 | 0.4251 |
1328
+ | **cosine_ndcg@10** | **0.4992** |
1329
+ | cosine_mrr@10 | 0.483 |
1330
+ | cosine_map@100 | 0.6025 |
1331
+
1332
+ #### Information Retrieval
1333
+
1334
+ * Dataset: `dim_64`
1335
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1336
+ ```json
1337
+ {
1338
+ "truncate_dim": 64
1339
+ }
1340
+ ```
1341
+
1342
+ | Metric | Value |
1343
+ |:--------------------|:-----------|
1344
+ | cosine_accuracy@1 | 0.4762 |
1345
+ | cosine_accuracy@3 | 0.4762 |
1346
+ | cosine_accuracy@5 | 0.4762 |
1347
+ | cosine_accuracy@10 | 0.5238 |
1348
+ | cosine_precision@1 | 0.4762 |
1349
+ | cosine_precision@3 | 0.4603 |
1350
+ | cosine_precision@5 | 0.419 |
1351
+ | cosine_precision@10 | 0.3762 |
1352
+ | cosine_recall@1 | 0.0782 |
1353
+ | cosine_recall@3 | 0.2109 |
1354
+ | cosine_recall@5 | 0.276 |
1355
+ | cosine_recall@10 | 0.4251 |
1356
+ | **cosine_ndcg@10** | **0.4992** |
1357
+ | cosine_mrr@10 | 0.483 |
1358
+ | cosine_map@100 | 0.596 |
1359
+
1360
+ <!--
1361
+ ## Bias, Risks and Limitations
1362
+
1363
+ *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
1364
+ -->
1365
+
1366
+ <!--
1367
+ ### Recommendations
1368
+
1369
+ *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
1370
+ -->
1371
+
1372
+ ## Training Details
1373
+
1374
+ ### Training Dataset
1375
+
1376
+ #### Unnamed Dataset
1377
+
1378
+ * Size: 82 training samples
1379
+ * Columns: <code>anchor</code> and <code>positive</code>
1380
+ * Approximate statistics based on the first 82 samples:
1381
+ | | anchor | positive |
1382
+ |:--------|:----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|
1383
+ | type | string | string |
1384
+ | details | <ul><li>min: 9 tokens</li><li>mean: 18.17 tokens</li><li>max: 34 tokens</li></ul> | <ul><li>min: 69 tokens</li><li>mean: 399.51 tokens</li><li>max: 512 tokens</li></ul> |
1385
+ * Samples:
1386
+ | anchor | positive |
1387
+ |:----------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
1388
+ | <code>What determines whether the act in question shall be punished if the offender is in the service of the legal holder of the data?</code> | <code>Everyone who obtains access to data recorded in a computer or in the external memory of a computer or transmitted by telecommunication systems shall be punished with imprisonment for up to six months or by a fine from 29 to 15,000 Euro, under the condition that these acts have been committed without right, especially in violation of prohibitions or of security measures taken by the legal holder. If the act concerns the international relations or the security of the State, he shall be punished according to Article 148.<br>If the offender is in the service of the legal holder of the data, the act of the preceding paragraph shall be punished only if it has been explicitly prohibited by internal regulations or by a written decision of the holder or of a competent employee of his.<br></code> |
1389
+ | <code>What must be causally connected to the perpetrator's deceptive acts?</code> | <code>According to Article 386 paragraph 1 of the Greek Penal Code,<br><br>"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."<br><br>From these provisions, it follows that, for the crime of fraud to be established, the following elements are required:<br><br>a) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit;<br><br>b) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence detrimental to th...</code> |
1390
+ | <code>Who can be punished with imprisonment?</code> | <code>1. Anyone who, by knowingly presenting false facts as true or by unlawfully concealing or withholding true facts, damages another person's property by persuading someone to act, omission, or tolerance with the aim of obtaining, for themselves or another, an unlawful financial gain from the damage to that property shall be punished with imprisonment, "and if the damage caused is particularly great, with imprisonment of at least three (3) months and a fine." .<br>If the damage caused exceeds a total of one hundred and twenty thousand (120,000) euros, imprisonment of up to ten (10) years and a fine shall be imposed.<br>2. If the fraud is directed directly against the legal entity of the Greek State, legal entities governed by public law, or local government organizations, and the damage caused exceeds a total of one hundred and twenty thousand (120,000) euros, a prison sentence of at least ten (10) years and a fine of up to one thousand (1,000) daily units shall be imposed. This offense shall b...</code> |
1391
+ * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters:
1392
+ ```json
1393
+ {
1394
+ "loss": "MultipleNegativesRankingLoss",
1395
+ "matryoshka_dims": [
1396
+ 1024,
1397
+ 768,
1398
+ 512,
1399
+ 256,
1400
+ 128,
1401
+ 64
1402
+ ],
1403
+ "matryoshka_weights": [
1404
+ 1,
1405
+ 1,
1406
+ 1,
1407
+ 1,
1408
+ 1,
1409
+ 1
1410
+ ],
1411
+ "n_dims_per_step": -1
1412
+ }
1413
+ ```
1414
+
1415
+ ### Training Hyperparameters
1416
+ #### Non-Default Hyperparameters
1417
+
1418
+ - `eval_strategy`: epoch
1419
+ - `gradient_accumulation_steps`: 2
1420
+ - `learning_rate`: 2e-05
1421
+ - `num_train_epochs`: 10
1422
+ - `lr_scheduler_type`: cosine
1423
+ - `warmup_ratio`: 0.1
1424
+ - `bf16`: True
1425
+ - `tf32`: True
1426
+ - `load_best_model_at_end`: True
1427
+ - `optim`: adamw_torch_fused
1428
+ - `batch_sampler`: no_duplicates
1429
+
1430
+ #### All Hyperparameters
1431
+ <details><summary>Click to expand</summary>
1432
+
1433
+ - `overwrite_output_dir`: False
1434
+ - `do_predict`: False
1435
+ - `eval_strategy`: epoch
1436
+ - `prediction_loss_only`: True
1437
+ - `per_device_train_batch_size`: 8
1438
+ - `per_device_eval_batch_size`: 8
1439
+ - `per_gpu_train_batch_size`: None
1440
+ - `per_gpu_eval_batch_size`: None
1441
+ - `gradient_accumulation_steps`: 2
1442
+ - `eval_accumulation_steps`: None
1443
+ - `torch_empty_cache_steps`: None
1444
+ - `learning_rate`: 2e-05
1445
+ - `weight_decay`: 0.0
1446
+ - `adam_beta1`: 0.9
1447
+ - `adam_beta2`: 0.999
1448
+ - `adam_epsilon`: 1e-08
1449
+ - `max_grad_norm`: 1.0
1450
+ - `num_train_epochs`: 10
1451
+ - `max_steps`: -1
1452
+ - `lr_scheduler_type`: cosine
1453
+ - `lr_scheduler_kwargs`: {}
1454
+ - `warmup_ratio`: 0.1
1455
+ - `warmup_steps`: 0
1456
+ - `log_level`: passive
1457
+ - `log_level_replica`: warning
1458
+ - `log_on_each_node`: True
1459
+ - `logging_nan_inf_filter`: True
1460
+ - `save_safetensors`: True
1461
+ - `save_on_each_node`: False
1462
+ - `save_only_model`: False
1463
+ - `restore_callback_states_from_checkpoint`: False
1464
+ - `no_cuda`: False
1465
+ - `use_cpu`: False
1466
+ - `use_mps_device`: False
1467
+ - `seed`: 42
1468
+ - `data_seed`: None
1469
+ - `jit_mode_eval`: False
1470
+ - `use_ipex`: False
1471
+ - `bf16`: True
1472
+ - `fp16`: False
1473
+ - `fp16_opt_level`: O1
1474
+ - `half_precision_backend`: auto
1475
+ - `bf16_full_eval`: False
1476
+ - `fp16_full_eval`: False
1477
+ - `tf32`: True
1478
+ - `local_rank`: 0
1479
+ - `ddp_backend`: None
1480
+ - `tpu_num_cores`: None
1481
+ - `tpu_metrics_debug`: False
1482
+ - `debug`: []
1483
+ - `dataloader_drop_last`: False
1484
+ - `dataloader_num_workers`: 0
1485
+ - `dataloader_prefetch_factor`: None
1486
+ - `past_index`: -1
1487
+ - `disable_tqdm`: False
1488
+ - `remove_unused_columns`: True
1489
+ - `label_names`: None
1490
+ - `load_best_model_at_end`: True
1491
+ - `ignore_data_skip`: False
1492
+ - `fsdp`: []
1493
+ - `fsdp_min_num_params`: 0
1494
+ - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
1495
+ - `tp_size`: 0
1496
+ - `fsdp_transformer_layer_cls_to_wrap`: None
1497
+ - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
1498
+ - `deepspeed`: None
1499
+ - `label_smoothing_factor`: 0.0
1500
+ - `optim`: adamw_torch_fused
1501
+ - `optim_args`: None
1502
+ - `adafactor`: False
1503
+ - `group_by_length`: False
1504
+ - `length_column_name`: length
1505
+ - `ddp_find_unused_parameters`: None
1506
+ - `ddp_bucket_cap_mb`: None
1507
+ - `ddp_broadcast_buffers`: False
1508
+ - `dataloader_pin_memory`: True
1509
+ - `dataloader_persistent_workers`: False
1510
+ - `skip_memory_metrics`: True
1511
+ - `use_legacy_prediction_loop`: False
1512
+ - `push_to_hub`: False
1513
+ - `resume_from_checkpoint`: None
1514
+ - `hub_model_id`: None
1515
+ - `hub_strategy`: every_save
1516
+ - `hub_private_repo`: None
1517
+ - `hub_always_push`: False
1518
+ - `gradient_checkpointing`: False
1519
+ - `gradient_checkpointing_kwargs`: None
1520
+ - `include_inputs_for_metrics`: False
1521
+ - `include_for_metrics`: []
1522
+ - `eval_do_concat_batches`: True
1523
+ - `fp16_backend`: auto
1524
+ - `push_to_hub_model_id`: None
1525
+ - `push_to_hub_organization`: None
1526
+ - `mp_parameters`:
1527
+ - `auto_find_batch_size`: False
1528
+ - `full_determinism`: False
1529
+ - `torchdynamo`: None
1530
+ - `ray_scope`: last
1531
+ - `ddp_timeout`: 1800
1532
+ - `torch_compile`: False
1533
+ - `torch_compile_backend`: None
1534
+ - `torch_compile_mode`: None
1535
+ - `include_tokens_per_second`: False
1536
+ - `include_num_input_tokens_seen`: False
1537
+ - `neftune_noise_alpha`: None
1538
+ - `optim_target_modules`: None
1539
+ - `batch_eval_metrics`: False
1540
+ - `eval_on_start`: False
1541
+ - `use_liger_kernel`: False
1542
+ - `eval_use_gather_object`: False
1543
+ - `average_tokens_across_devices`: False
1544
+ - `prompts`: None
1545
+ - `batch_sampler`: no_duplicates
1546
+ - `multi_dataset_batch_sampler`: proportional
1547
+ - `router_mapping`: {}
1548
+ - `learning_rate_mapping`: {}
1549
+
1550
+ </details>
1551
+
1552
+ ### Training Logs
1553
+ | Epoch | Step | Training Loss | dim_1024_cosine_ndcg@10 | dim_768_cosine_ndcg@10 | dim_512_cosine_ndcg@10 | dim_256_cosine_ndcg@10 | dim_128_cosine_ndcg@10 | dim_64_cosine_ndcg@10 |
1554
+ |:------:|:----:|:-------------:|:-----------------------:|:----------------------:|:----------------------:|:----------------------:|:----------------------:|:---------------------:|
1555
+ | 0.1818 | 1 | 18.029 | - | - | - | - | - | - |
1556
+ | 0.3636 | 2 | 19.4106 | - | - | - | - | - | - |
1557
+ | 0.5455 | 3 | 16.6201 | - | - | - | - | - | - |
1558
+ | 0.7273 | 4 | 15.3048 | - | - | - | - | - | - |
1559
+ | 0.9091 | 5 | 14.0182 | - | - | - | - | - | - |
1560
+ | 1.0 | 6 | 6.4771 | - | - | - | - | - | - |
1561
+ | 1.0909 | 7 | 6.7664 | 0.6167 | 0.5821 | 0.5524 | 0.5177 | 0.5278 | 0.4124 |
1562
+ | 1.1818 | 8 | 11.8583 | - | - | - | - | - | - |
1563
+ | 1.3636 | 9 | 11.9216 | - | - | - | - | - | - |
1564
+ | 1.5455 | 10 | 13.3764 | - | - | - | - | - | - |
1565
+ | 1.7273 | 11 | 12.9063 | - | - | - | - | - | - |
1566
+ | 1.9091 | 12 | 13.5984 | - | - | - | - | - | - |
1567
+ | 2.0 | 13 | 7.8523 | - | - | - | - | - | - |
1568
+ | 2.0909 | 14 | 4.4487 | 0.5921 | 0.5921 | 0.5518 | 0.5709 | 0.5685 | 0.5113 |
1569
+ | 2.1818 | 15 | 8.5374 | - | - | - | - | - | - |
1570
+ | 2.3636 | 16 | 9.6999 | - | - | - | - | - | - |
1571
+ | 2.5455 | 17 | 9.0121 | - | - | - | - | - | - |
1572
+ | 2.7273 | 18 | 13.5705 | - | - | - | - | - | - |
1573
+ | 2.9091 | 19 | 13.0195 | - | - | - | - | - | - |
1574
+ | 3.0 | 20 | 7.9821 | - | - | - | - | - | - |
1575
+ | 3.0909 | 21 | 3.2842 | 0.5159 | 0.5636 | 0.5468 | 0.5468 | 0.5468 | 0.5233 |
1576
+ | 3.1818 | 22 | 4.4446 | - | - | - | - | - | - |
1577
+ | 3.3636 | 23 | 5.7244 | - | - | - | - | - | - |
1578
+ | 3.5455 | 24 | 7.1394 | - | - | - | - | - | - |
1579
+ | 3.7273 | 25 | 16.7583 | - | - | - | - | - | - |
1580
+ | 3.9091 | 26 | 11.3515 | - | - | - | - | - | - |
1581
+ | 4.0 | 27 | 8.813 | - | - | - | - | - | - |
1582
+ | 4.0909 | 28 | 6.9124 | 0.5159 | 0.5468 | 0.4992 | 0.5468 | 0.4992 | 0.4992 |
1583
+
1584
+
1585
+ ### Framework Versions
1586
+ - Python: 3.12.12
1587
+ - Sentence Transformers: 5.1.1
1588
+ - Transformers: 4.51.3
1589
+ - PyTorch: 2.8.0+cu126
1590
+ - Accelerate: 1.11.0
1591
+ - Datasets: 4.0.0
1592
+ - Tokenizers: 0.21.4
1593
+
1594
+ ## Citation
1595
+
1596
+ ### BibTeX
1597
+
1598
+ #### Sentence Transformers
1599
+ ```bibtex
1600
+ @inproceedings{reimers-2019-sentence-bert,
1601
+ title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
1602
+ author = "Reimers, Nils and Gurevych, Iryna",
1603
+ booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
1604
+ month = "11",
1605
+ year = "2019",
1606
+ publisher = "Association for Computational Linguistics",
1607
+ url = "https://arxiv.org/abs/1908.10084",
1608
+ }
1609
+ ```
1610
+
1611
+ #### MatryoshkaLoss
1612
+ ```bibtex
1613
+ @misc{kusupati2024matryoshka,
1614
+ title={Matryoshka Representation Learning},
1615
+ 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},
1616
+ year={2024},
1617
+ eprint={2205.13147},
1618
+ archivePrefix={arXiv},
1619
+ primaryClass={cs.LG}
1620
+ }
1621
+ ```
1622
+
1623
+ #### MultipleNegativesRankingLoss
1624
+ ```bibtex
1625
+ @misc{henderson2017efficient,
1626
+ title={Efficient Natural Language Response Suggestion for Smart Reply},
1627
+ 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},
1628
+ year={2017},
1629
+ eprint={1705.00652},
1630
+ archivePrefix={arXiv},
1631
+ primaryClass={cs.CL}
1632
+ }
1633
+ ```
1634
+
1635
+ <!--
1636
+ ## Glossary
1637
+
1638
+ *Clearly define terms in order to be accessible across audiences.*
1639
+ -->
1640
+
1641
+ <!--
1642
+ ## Model Card Authors
1643
+
1644
+ *Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
1645
+ -->
1646
+
1647
+ <!--
1648
+ ## Model Card Contact
1649
+
1650
+ *Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
1651
+ -->
checkpoint-28/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "XLMRobertaModel"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 0,
7
+ "classifier_dropout": null,
8
+ "eos_token_id": 2,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 1024,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 4096,
14
+ "layer_norm_eps": 1e-05,
15
+ "max_position_embeddings": 514,
16
+ "model_type": "xlm-roberta",
17
+ "num_attention_heads": 16,
18
+ "num_hidden_layers": 24,
19
+ "output_past": true,
20
+ "pad_token_id": 1,
21
+ "position_embedding_type": "absolute",
22
+ "torch_dtype": "float32",
23
+ "transformers_version": "4.51.3",
24
+ "type_vocab_size": 1,
25
+ "use_cache": true,
26
+ "vocab_size": 250002
27
+ }
checkpoint-28/config_sentence_transformers.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "SentenceTransformer",
3
+ "__version__": {
4
+ "sentence_transformers": "5.1.1",
5
+ "transformers": "4.51.3",
6
+ "pytorch": "2.8.0+cu126"
7
+ },
8
+ "prompts": {
9
+ "query": "",
10
+ "document": ""
11
+ },
12
+ "default_prompt_name": null,
13
+ "similarity_fn_name": "cosine"
14
+ }
checkpoint-28/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f97fcbc0e7a951f68c8432dcdd5f2ecd96d044d3c202e0790218b87ffa9a91d
3
+ size 2239607176
checkpoint-28/modules.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "idx": 0,
4
+ "name": "0",
5
+ "path": "",
6
+ "type": "sentence_transformers.models.Transformer"
7
+ },
8
+ {
9
+ "idx": 1,
10
+ "name": "1",
11
+ "path": "1_Pooling",
12
+ "type": "sentence_transformers.models.Pooling"
13
+ },
14
+ {
15
+ "idx": 2,
16
+ "name": "2",
17
+ "path": "2_Normalize",
18
+ "type": "sentence_transformers.models.Normalize"
19
+ }
20
+ ]
checkpoint-28/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dc4a1e358287ce2ca3e23fd5b7842322992a275fb872bc5b97cea500ed439d2c
3
+ size 4471067142
checkpoint-28/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3171c785711284de5dd361b74d324b51f22ae04ab6c679c768dfe351168dbec4
3
+ size 14645
checkpoint-28/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c2b4fb22f017dc0d2e21c4ae1cf21c451d5bfd5c49a6b6e315bfff5d4ae75a1d
3
+ size 1465
checkpoint-28/sentence_bert_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "max_seq_length": 512,
3
+ "do_lower_case": false
4
+ }
checkpoint-28/sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
3
+ size 5069051
checkpoint-28/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
checkpoint-28/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:883b037111086fd4dfebbbc9b7cee11e1517b5e0c0514879478661440f137085
3
+ size 17082987
checkpoint-28/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": true,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
checkpoint-28/trainer_state.json ADDED
@@ -0,0 +1,631 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 14,
3
+ "best_metric": 0.5685354415901852,
4
+ "best_model_checkpoint": "multilingual-e5-large/checkpoint-14",
5
+ "epoch": 4.090909090909091,
6
+ "eval_steps": 500,
7
+ "global_step": 28,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.18181818181818182,
14
+ "grad_norm": 135.6220703125,
15
+ "learning_rate": 0.0,
16
+ "loss": 18.029,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.36363636363636365,
21
+ "grad_norm": 140.5317840576172,
22
+ "learning_rate": 4.000000000000001e-06,
23
+ "loss": 19.4106,
24
+ "step": 2
25
+ },
26
+ {
27
+ "epoch": 0.5454545454545454,
28
+ "grad_norm": 142.2399444580078,
29
+ "learning_rate": 8.000000000000001e-06,
30
+ "loss": 16.6201,
31
+ "step": 3
32
+ },
33
+ {
34
+ "epoch": 0.7272727272727273,
35
+ "grad_norm": 110.63945770263672,
36
+ "learning_rate": 1.2e-05,
37
+ "loss": 15.3048,
38
+ "step": 4
39
+ },
40
+ {
41
+ "epoch": 0.9090909090909091,
42
+ "grad_norm": 129.43089294433594,
43
+ "learning_rate": 1.6000000000000003e-05,
44
+ "loss": 14.0182,
45
+ "step": 5
46
+ },
47
+ {
48
+ "epoch": 1.0,
49
+ "grad_norm": 108.25520324707031,
50
+ "learning_rate": 2e-05,
51
+ "loss": 6.4771,
52
+ "step": 6
53
+ },
54
+ {
55
+ "epoch": 1.0909090909090908,
56
+ "grad_norm": 95.05782318115234,
57
+ "learning_rate": 1.9975640502598243e-05,
58
+ "loss": 6.7664,
59
+ "step": 7
60
+ },
61
+ {
62
+ "epoch": 1.0909090909090908,
63
+ "eval_dim_1024_cosine_accuracy@1": 0.5714285714285714,
64
+ "eval_dim_1024_cosine_accuracy@10": 0.6666666666666666,
65
+ "eval_dim_1024_cosine_accuracy@3": 0.5714285714285714,
66
+ "eval_dim_1024_cosine_accuracy@5": 0.6190476190476191,
67
+ "eval_dim_1024_cosine_map@100": 0.6946498352795571,
68
+ "eval_dim_1024_cosine_mrr@10": 0.5888888888888888,
69
+ "eval_dim_1024_cosine_ndcg@10": 0.6167053425672016,
70
+ "eval_dim_1024_cosine_precision@1": 0.5714285714285714,
71
+ "eval_dim_1024_cosine_precision@10": 0.4476190476190477,
72
+ "eval_dim_1024_cosine_precision@3": 0.5555555555555556,
73
+ "eval_dim_1024_cosine_precision@5": 0.5142857142857142,
74
+ "eval_dim_1024_cosine_recall@1": 0.10123626373626372,
75
+ "eval_dim_1024_cosine_recall@10": 0.5401404151404151,
76
+ "eval_dim_1024_cosine_recall@3": 0.2798992673992674,
77
+ "eval_dim_1024_cosine_recall@5": 0.3871336996336997,
78
+ "eval_dim_128_cosine_accuracy@1": 0.47619047619047616,
79
+ "eval_dim_128_cosine_accuracy@10": 0.5714285714285714,
80
+ "eval_dim_128_cosine_accuracy@3": 0.47619047619047616,
81
+ "eval_dim_128_cosine_accuracy@5": 0.5238095238095238,
82
+ "eval_dim_128_cosine_map@100": 0.5954946117864774,
83
+ "eval_dim_128_cosine_mrr@10": 0.49251700680272104,
84
+ "eval_dim_128_cosine_ndcg@10": 0.5277801377642622,
85
+ "eval_dim_128_cosine_precision@1": 0.47619047619047616,
86
+ "eval_dim_128_cosine_precision@10": 0.3761904761904762,
87
+ "eval_dim_128_cosine_precision@3": 0.4603174603174603,
88
+ "eval_dim_128_cosine_precision@5": 0.42857142857142855,
89
+ "eval_dim_128_cosine_recall@1": 0.08536324786324785,
90
+ "eval_dim_128_cosine_recall@10": 0.4806166056166057,
91
+ "eval_dim_128_cosine_recall@3": 0.23228021978021982,
92
+ "eval_dim_128_cosine_recall@5": 0.31967338217338215,
93
+ "eval_dim_256_cosine_accuracy@1": 0.47619047619047616,
94
+ "eval_dim_256_cosine_accuracy@10": 0.5714285714285714,
95
+ "eval_dim_256_cosine_accuracy@3": 0.47619047619047616,
96
+ "eval_dim_256_cosine_accuracy@5": 0.5238095238095238,
97
+ "eval_dim_256_cosine_map@100": 0.6218563539505393,
98
+ "eval_dim_256_cosine_mrr@10": 0.4925170068027211,
99
+ "eval_dim_256_cosine_ndcg@10": 0.5177339964253599,
100
+ "eval_dim_256_cosine_precision@1": 0.47619047619047616,
101
+ "eval_dim_256_cosine_precision@10": 0.3476190476190476,
102
+ "eval_dim_256_cosine_precision@3": 0.4603174603174603,
103
+ "eval_dim_256_cosine_precision@5": 0.419047619047619,
104
+ "eval_dim_256_cosine_recall@1": 0.09543650793650793,
105
+ "eval_dim_256_cosine_recall@10": 0.47817460317460325,
106
+ "eval_dim_256_cosine_recall@3": 0.2625,
107
+ "eval_dim_256_cosine_recall@5": 0.35813492063492064,
108
+ "eval_dim_512_cosine_accuracy@1": 0.5238095238095238,
109
+ "eval_dim_512_cosine_accuracy@10": 0.5714285714285714,
110
+ "eval_dim_512_cosine_accuracy@3": 0.5238095238095238,
111
+ "eval_dim_512_cosine_accuracy@5": 0.5714285714285714,
112
+ "eval_dim_512_cosine_map@100": 0.6608463504287858,
113
+ "eval_dim_512_cosine_mrr@10": 0.5333333333333333,
114
+ "eval_dim_512_cosine_ndcg@10": 0.5523687509163372,
115
+ "eval_dim_512_cosine_precision@1": 0.5238095238095238,
116
+ "eval_dim_512_cosine_precision@10": 0.3761904761904762,
117
+ "eval_dim_512_cosine_precision@3": 0.5079365079365079,
118
+ "eval_dim_512_cosine_precision@5": 0.4666666666666666,
119
+ "eval_dim_512_cosine_recall@1": 0.0994047619047619,
120
+ "eval_dim_512_cosine_recall@10": 0.5019841269841271,
121
+ "eval_dim_512_cosine_recall@3": 0.2744047619047619,
122
+ "eval_dim_512_cosine_recall@5": 0.37797619047619047,
123
+ "eval_dim_64_cosine_accuracy@1": 0.3333333333333333,
124
+ "eval_dim_64_cosine_accuracy@10": 0.5714285714285714,
125
+ "eval_dim_64_cosine_accuracy@3": 0.38095238095238093,
126
+ "eval_dim_64_cosine_accuracy@5": 0.38095238095238093,
127
+ "eval_dim_64_cosine_map@100": 0.5579595177809107,
128
+ "eval_dim_64_cosine_mrr@10": 0.3764172335600907,
129
+ "eval_dim_64_cosine_ndcg@10": 0.41244392103396355,
130
+ "eval_dim_64_cosine_precision@1": 0.3333333333333333,
131
+ "eval_dim_64_cosine_precision@10": 0.2619047619047619,
132
+ "eval_dim_64_cosine_precision@3": 0.3492063492063492,
133
+ "eval_dim_64_cosine_precision@5": 0.3142857142857143,
134
+ "eval_dim_64_cosine_recall@1": 0.07063492063492063,
135
+ "eval_dim_64_cosine_recall@10": 0.42002442002442003,
136
+ "eval_dim_64_cosine_recall@3": 0.2357142857142857,
137
+ "eval_dim_64_cosine_recall@5": 0.34523809523809523,
138
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
139
+ "eval_dim_768_cosine_accuracy@10": 0.6666666666666666,
140
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
141
+ "eval_dim_768_cosine_accuracy@5": 0.5714285714285714,
142
+ "eval_dim_768_cosine_map@100": 0.6727109022414055,
143
+ "eval_dim_768_cosine_mrr@10": 0.5480725623582765,
144
+ "eval_dim_768_cosine_ndcg@10": 0.5820705880762242,
145
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
146
+ "eval_dim_768_cosine_precision@10": 0.41904761904761906,
147
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
148
+ "eval_dim_768_cosine_precision@5": 0.4666666666666666,
149
+ "eval_dim_768_cosine_recall@1": 0.09726800976800977,
150
+ "eval_dim_768_cosine_recall@10": 0.5163308913308913,
151
+ "eval_dim_768_cosine_recall@3": 0.2679945054945055,
152
+ "eval_dim_768_cosine_recall@5": 0.36729242979242976,
153
+ "eval_runtime": 7.0357,
154
+ "eval_samples_per_second": 0.0,
155
+ "eval_sequential_score": 0.41244392103396355,
156
+ "eval_steps_per_second": 0.0,
157
+ "step": 7
158
+ },
159
+ {
160
+ "epoch": 1.1818181818181819,
161
+ "grad_norm": 159.73828125,
162
+ "learning_rate": 1.9902680687415704e-05,
163
+ "loss": 11.8583,
164
+ "step": 8
165
+ },
166
+ {
167
+ "epoch": 1.3636363636363638,
168
+ "grad_norm": 134.70603942871094,
169
+ "learning_rate": 1.9781476007338058e-05,
170
+ "loss": 11.9216,
171
+ "step": 9
172
+ },
173
+ {
174
+ "epoch": 1.5454545454545454,
175
+ "grad_norm": 139.73654174804688,
176
+ "learning_rate": 1.961261695938319e-05,
177
+ "loss": 13.3764,
178
+ "step": 10
179
+ },
180
+ {
181
+ "epoch": 1.7272727272727273,
182
+ "grad_norm": 173.23065185546875,
183
+ "learning_rate": 1.9396926207859085e-05,
184
+ "loss": 12.9063,
185
+ "step": 11
186
+ },
187
+ {
188
+ "epoch": 1.9090909090909092,
189
+ "grad_norm": 208.78482055664062,
190
+ "learning_rate": 1.913545457642601e-05,
191
+ "loss": 13.5984,
192
+ "step": 12
193
+ },
194
+ {
195
+ "epoch": 2.0,
196
+ "grad_norm": 298.82183837890625,
197
+ "learning_rate": 1.8829475928589272e-05,
198
+ "loss": 7.8523,
199
+ "step": 13
200
+ },
201
+ {
202
+ "epoch": 2.090909090909091,
203
+ "grad_norm": 156.5718994140625,
204
+ "learning_rate": 1.848048096156426e-05,
205
+ "loss": 4.4487,
206
+ "step": 14
207
+ },
208
+ {
209
+ "epoch": 2.090909090909091,
210
+ "eval_dim_1024_cosine_accuracy@1": 0.5238095238095238,
211
+ "eval_dim_1024_cosine_accuracy@10": 0.6666666666666666,
212
+ "eval_dim_1024_cosine_accuracy@3": 0.5238095238095238,
213
+ "eval_dim_1024_cosine_accuracy@5": 0.5714285714285714,
214
+ "eval_dim_1024_cosine_map@100": 0.67423207909377,
215
+ "eval_dim_1024_cosine_mrr@10": 0.5480725623582765,
216
+ "eval_dim_1024_cosine_ndcg@10": 0.5921167294151266,
217
+ "eval_dim_1024_cosine_precision@1": 0.5238095238095238,
218
+ "eval_dim_1024_cosine_precision@10": 0.4476190476190477,
219
+ "eval_dim_1024_cosine_precision@3": 0.5079365079365079,
220
+ "eval_dim_1024_cosine_precision@5": 0.47619047619047616,
221
+ "eval_dim_1024_cosine_recall@1": 0.08933150183150182,
222
+ "eval_dim_1024_cosine_recall@10": 0.5401404151404151,
223
+ "eval_dim_1024_cosine_recall@3": 0.24418498168498168,
224
+ "eval_dim_1024_cosine_recall@5": 0.33951465201465203,
225
+ "eval_dim_128_cosine_accuracy@1": 0.5238095238095238,
226
+ "eval_dim_128_cosine_accuracy@10": 0.6190476190476191,
227
+ "eval_dim_128_cosine_accuracy@3": 0.5238095238095238,
228
+ "eval_dim_128_cosine_accuracy@5": 0.5714285714285714,
229
+ "eval_dim_128_cosine_map@100": 0.6489604480560528,
230
+ "eval_dim_128_cosine_mrr@10": 0.5401360544217686,
231
+ "eval_dim_128_cosine_ndcg@10": 0.5685354415901852,
232
+ "eval_dim_128_cosine_precision@1": 0.5238095238095238,
233
+ "eval_dim_128_cosine_precision@10": 0.4238095238095239,
234
+ "eval_dim_128_cosine_precision@3": 0.5238095238095238,
235
+ "eval_dim_128_cosine_precision@5": 0.5047619047619047,
236
+ "eval_dim_128_cosine_recall@1": 0.07345848595848595,
237
+ "eval_dim_128_cosine_recall@10": 0.5202991452991453,
238
+ "eval_dim_128_cosine_recall@3": 0.2203754578754579,
239
+ "eval_dim_128_cosine_recall@5": 0.34745115995116,
240
+ "eval_dim_256_cosine_accuracy@1": 0.5238095238095238,
241
+ "eval_dim_256_cosine_accuracy@10": 0.6190476190476191,
242
+ "eval_dim_256_cosine_accuracy@3": 0.5238095238095238,
243
+ "eval_dim_256_cosine_accuracy@5": 0.5714285714285714,
244
+ "eval_dim_256_cosine_map@100": 0.651530364911684,
245
+ "eval_dim_256_cosine_mrr@10": 0.5401360544217686,
246
+ "eval_dim_256_cosine_ndcg@10": 0.5708936958722651,
247
+ "eval_dim_256_cosine_precision@1": 0.5238095238095238,
248
+ "eval_dim_256_cosine_precision@10": 0.4238095238095239,
249
+ "eval_dim_256_cosine_precision@3": 0.5079365079365079,
250
+ "eval_dim_256_cosine_precision@5": 0.49523809523809514,
251
+ "eval_dim_256_cosine_recall@1": 0.0813949938949939,
252
+ "eval_dim_256_cosine_recall@10": 0.5202991452991453,
253
+ "eval_dim_256_cosine_recall@3": 0.22037545787545787,
254
+ "eval_dim_256_cosine_recall@5": 0.33951465201465203,
255
+ "eval_dim_512_cosine_accuracy@1": 0.47619047619047616,
256
+ "eval_dim_512_cosine_accuracy@10": 0.6190476190476191,
257
+ "eval_dim_512_cosine_accuracy@3": 0.47619047619047616,
258
+ "eval_dim_512_cosine_accuracy@5": 0.5714285714285714,
259
+ "eval_dim_512_cosine_map@100": 0.6265911712939339,
260
+ "eval_dim_512_cosine_mrr@10": 0.5020408163265305,
261
+ "eval_dim_512_cosine_ndcg@10": 0.5518338753600308,
262
+ "eval_dim_512_cosine_precision@1": 0.47619047619047616,
263
+ "eval_dim_512_cosine_precision@10": 0.4238095238095239,
264
+ "eval_dim_512_cosine_precision@3": 0.4603174603174603,
265
+ "eval_dim_512_cosine_precision@5": 0.45714285714285713,
266
+ "eval_dim_512_cosine_recall@1": 0.07345848595848595,
267
+ "eval_dim_512_cosine_recall@10": 0.5202991452991453,
268
+ "eval_dim_512_cosine_recall@3": 0.19656593406593406,
269
+ "eval_dim_512_cosine_recall@5": 0.3077686202686203,
270
+ "eval_dim_64_cosine_accuracy@1": 0.42857142857142855,
271
+ "eval_dim_64_cosine_accuracy@10": 0.6190476190476191,
272
+ "eval_dim_64_cosine_accuracy@3": 0.42857142857142855,
273
+ "eval_dim_64_cosine_accuracy@5": 0.47619047619047616,
274
+ "eval_dim_64_cosine_map@100": 0.5888462989137369,
275
+ "eval_dim_64_cosine_mrr@10": 0.45963718820861665,
276
+ "eval_dim_64_cosine_ndcg@10": 0.51131642091388,
277
+ "eval_dim_64_cosine_precision@1": 0.42857142857142855,
278
+ "eval_dim_64_cosine_precision@10": 0.3999999999999999,
279
+ "eval_dim_64_cosine_precision@3": 0.42857142857142855,
280
+ "eval_dim_64_cosine_precision@5": 0.42857142857142855,
281
+ "eval_dim_64_cosine_recall@1": 0.053617216117216114,
282
+ "eval_dim_64_cosine_recall@10": 0.5004578754578755,
283
+ "eval_dim_64_cosine_recall@3": 0.16085164835164836,
284
+ "eval_dim_64_cosine_recall@5": 0.27205433455433453,
285
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
286
+ "eval_dim_768_cosine_accuracy@10": 0.6666666666666666,
287
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
288
+ "eval_dim_768_cosine_accuracy@5": 0.5714285714285714,
289
+ "eval_dim_768_cosine_map@100": 0.67423207909377,
290
+ "eval_dim_768_cosine_mrr@10": 0.5480725623582765,
291
+ "eval_dim_768_cosine_ndcg@10": 0.5921167294151266,
292
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
293
+ "eval_dim_768_cosine_precision@10": 0.4476190476190477,
294
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
295
+ "eval_dim_768_cosine_precision@5": 0.47619047619047616,
296
+ "eval_dim_768_cosine_recall@1": 0.08933150183150182,
297
+ "eval_dim_768_cosine_recall@10": 0.5401404151404151,
298
+ "eval_dim_768_cosine_recall@3": 0.24418498168498168,
299
+ "eval_dim_768_cosine_recall@5": 0.33951465201465203,
300
+ "eval_runtime": 6.9723,
301
+ "eval_samples_per_second": 0.0,
302
+ "eval_sequential_score": 0.51131642091388,
303
+ "eval_steps_per_second": 0.0,
304
+ "step": 14
305
+ },
306
+ {
307
+ "epoch": 2.1818181818181817,
308
+ "grad_norm": 156.31887817382812,
309
+ "learning_rate": 1.8090169943749477e-05,
310
+ "loss": 8.5374,
311
+ "step": 15
312
+ },
313
+ {
314
+ "epoch": 2.3636363636363638,
315
+ "grad_norm": 205.52923583984375,
316
+ "learning_rate": 1.766044443118978e-05,
317
+ "loss": 9.6999,
318
+ "step": 16
319
+ },
320
+ {
321
+ "epoch": 2.5454545454545454,
322
+ "grad_norm": 254.7772216796875,
323
+ "learning_rate": 1.7193398003386514e-05,
324
+ "loss": 9.0121,
325
+ "step": 17
326
+ },
327
+ {
328
+ "epoch": 2.7272727272727275,
329
+ "grad_norm": 476.75164794921875,
330
+ "learning_rate": 1.6691306063588583e-05,
331
+ "loss": 13.5705,
332
+ "step": 18
333
+ },
334
+ {
335
+ "epoch": 2.909090909090909,
336
+ "grad_norm": 376.09954833984375,
337
+ "learning_rate": 1.6156614753256583e-05,
338
+ "loss": 13.0195,
339
+ "step": 19
340
+ },
341
+ {
342
+ "epoch": 3.0,
343
+ "grad_norm": 387.24322509765625,
344
+ "learning_rate": 1.5591929034707468e-05,
345
+ "loss": 7.9821,
346
+ "step": 20
347
+ },
348
+ {
349
+ "epoch": 3.090909090909091,
350
+ "grad_norm": 210.42813110351562,
351
+ "learning_rate": 1.5000000000000002e-05,
352
+ "loss": 3.2842,
353
+ "step": 21
354
+ },
355
+ {
356
+ "epoch": 3.090909090909091,
357
+ "eval_dim_1024_cosine_accuracy@1": 0.47619047619047616,
358
+ "eval_dim_1024_cosine_accuracy@10": 0.5714285714285714,
359
+ "eval_dim_1024_cosine_accuracy@3": 0.47619047619047616,
360
+ "eval_dim_1024_cosine_accuracy@5": 0.47619047619047616,
361
+ "eval_dim_1024_cosine_map@100": 0.616370861587526,
362
+ "eval_dim_1024_cosine_mrr@10": 0.49092970521541945,
363
+ "eval_dim_1024_cosine_ndcg@10": 0.5159384546892658,
364
+ "eval_dim_1024_cosine_precision@1": 0.47619047619047616,
365
+ "eval_dim_1024_cosine_precision@10": 0.4,
366
+ "eval_dim_1024_cosine_precision@3": 0.4603174603174603,
367
+ "eval_dim_1024_cosine_precision@5": 0.419047619047619,
368
+ "eval_dim_1024_cosine_recall@1": 0.07822039072039072,
369
+ "eval_dim_1024_cosine_recall@10": 0.4449023199023199,
370
+ "eval_dim_1024_cosine_recall@3": 0.21085164835164832,
371
+ "eval_dim_1024_cosine_recall@5": 0.27602258852258854,
372
+ "eval_dim_128_cosine_accuracy@1": 0.5238095238095238,
373
+ "eval_dim_128_cosine_accuracy@10": 0.5714285714285714,
374
+ "eval_dim_128_cosine_accuracy@3": 0.5238095238095238,
375
+ "eval_dim_128_cosine_accuracy@5": 0.5238095238095238,
376
+ "eval_dim_128_cosine_map@100": 0.6278310491545506,
377
+ "eval_dim_128_cosine_mrr@10": 0.5306122448979591,
378
+ "eval_dim_128_cosine_ndcg@10": 0.5468399582764966,
379
+ "eval_dim_128_cosine_precision@1": 0.5238095238095238,
380
+ "eval_dim_128_cosine_precision@10": 0.4238095238095239,
381
+ "eval_dim_128_cosine_precision@3": 0.5079365079365079,
382
+ "eval_dim_128_cosine_precision@5": 0.4666666666666666,
383
+ "eval_dim_128_cosine_recall@1": 0.0811965811965812,
384
+ "eval_dim_128_cosine_recall@10": 0.4548229548229548,
385
+ "eval_dim_128_cosine_recall@3": 0.21978021978021975,
386
+ "eval_dim_128_cosine_recall@5": 0.2909035409035409,
387
+ "eval_dim_256_cosine_accuracy@1": 0.5238095238095238,
388
+ "eval_dim_256_cosine_accuracy@10": 0.5714285714285714,
389
+ "eval_dim_256_cosine_accuracy@3": 0.5238095238095238,
390
+ "eval_dim_256_cosine_accuracy@5": 0.5238095238095238,
391
+ "eval_dim_256_cosine_map@100": 0.6295752260160459,
392
+ "eval_dim_256_cosine_mrr@10": 0.5306122448979591,
393
+ "eval_dim_256_cosine_ndcg@10": 0.5468399582764966,
394
+ "eval_dim_256_cosine_precision@1": 0.5238095238095238,
395
+ "eval_dim_256_cosine_precision@10": 0.4238095238095239,
396
+ "eval_dim_256_cosine_precision@3": 0.5079365079365079,
397
+ "eval_dim_256_cosine_precision@5": 0.4666666666666666,
398
+ "eval_dim_256_cosine_recall@1": 0.0811965811965812,
399
+ "eval_dim_256_cosine_recall@10": 0.4548229548229548,
400
+ "eval_dim_256_cosine_recall@3": 0.21978021978021975,
401
+ "eval_dim_256_cosine_recall@5": 0.2909035409035409,
402
+ "eval_dim_512_cosine_accuracy@1": 0.5238095238095238,
403
+ "eval_dim_512_cosine_accuracy@10": 0.5714285714285714,
404
+ "eval_dim_512_cosine_accuracy@3": 0.5238095238095238,
405
+ "eval_dim_512_cosine_accuracy@5": 0.5238095238095238,
406
+ "eval_dim_512_cosine_map@100": 0.6298362122328386,
407
+ "eval_dim_512_cosine_mrr@10": 0.5306122448979591,
408
+ "eval_dim_512_cosine_ndcg@10": 0.5468399582764966,
409
+ "eval_dim_512_cosine_precision@1": 0.5238095238095238,
410
+ "eval_dim_512_cosine_precision@10": 0.4238095238095239,
411
+ "eval_dim_512_cosine_precision@3": 0.5079365079365079,
412
+ "eval_dim_512_cosine_precision@5": 0.4666666666666666,
413
+ "eval_dim_512_cosine_recall@1": 0.0811965811965812,
414
+ "eval_dim_512_cosine_recall@10": 0.4548229548229548,
415
+ "eval_dim_512_cosine_recall@3": 0.21978021978021975,
416
+ "eval_dim_512_cosine_recall@5": 0.2909035409035409,
417
+ "eval_dim_64_cosine_accuracy@1": 0.47619047619047616,
418
+ "eval_dim_64_cosine_accuracy@10": 0.5714285714285714,
419
+ "eval_dim_64_cosine_accuracy@3": 0.47619047619047616,
420
+ "eval_dim_64_cosine_accuracy@5": 0.5238095238095238,
421
+ "eval_dim_64_cosine_map@100": 0.6058793096030948,
422
+ "eval_dim_64_cosine_mrr@10": 0.49251700680272104,
423
+ "eval_dim_64_cosine_ndcg@10": 0.5232746482532176,
424
+ "eval_dim_64_cosine_precision@1": 0.47619047619047616,
425
+ "eval_dim_64_cosine_precision@10": 0.4238095238095238,
426
+ "eval_dim_64_cosine_precision@3": 0.4603174603174603,
427
+ "eval_dim_64_cosine_precision@5": 0.4476190476190475,
428
+ "eval_dim_64_cosine_recall@1": 0.06631562881562882,
429
+ "eval_dim_64_cosine_recall@10": 0.46474358974358976,
430
+ "eval_dim_64_cosine_recall@3": 0.17513736263736263,
431
+ "eval_dim_64_cosine_recall@5": 0.2641178266178266,
432
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
433
+ "eval_dim_768_cosine_accuracy@10": 0.6190476190476191,
434
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
435
+ "eval_dim_768_cosine_accuracy@5": 0.5238095238095238,
436
+ "eval_dim_768_cosine_map@100": 0.6486454426450273,
437
+ "eval_dim_768_cosine_mrr@10": 0.538548752834467,
438
+ "eval_dim_768_cosine_ndcg@10": 0.5635575023083134,
439
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
440
+ "eval_dim_768_cosine_precision@10": 0.4476190476190477,
441
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
442
+ "eval_dim_768_cosine_precision@5": 0.4666666666666666,
443
+ "eval_dim_768_cosine_recall@1": 0.0811965811965812,
444
+ "eval_dim_768_cosine_recall@10": 0.47466422466422464,
445
+ "eval_dim_768_cosine_recall@3": 0.21978021978021975,
446
+ "eval_dim_768_cosine_recall@5": 0.2909035409035409,
447
+ "eval_runtime": 7.0685,
448
+ "eval_samples_per_second": 0.0,
449
+ "eval_sequential_score": 0.5232746482532176,
450
+ "eval_steps_per_second": 0.0,
451
+ "step": 21
452
+ },
453
+ {
454
+ "epoch": 3.1818181818181817,
455
+ "grad_norm": 214.07061767578125,
456
+ "learning_rate": 1.4383711467890776e-05,
457
+ "loss": 4.4446,
458
+ "step": 22
459
+ },
460
+ {
461
+ "epoch": 3.3636363636363638,
462
+ "grad_norm": 234.80874633789062,
463
+ "learning_rate": 1.3746065934159123e-05,
464
+ "loss": 5.7244,
465
+ "step": 23
466
+ },
467
+ {
468
+ "epoch": 3.5454545454545454,
469
+ "grad_norm": 337.2349853515625,
470
+ "learning_rate": 1.3090169943749475e-05,
471
+ "loss": 7.1394,
472
+ "step": 24
473
+ },
474
+ {
475
+ "epoch": 3.7272727272727275,
476
+ "grad_norm": 562.9892578125,
477
+ "learning_rate": 1.2419218955996677e-05,
478
+ "loss": 16.7583,
479
+ "step": 25
480
+ },
481
+ {
482
+ "epoch": 3.909090909090909,
483
+ "grad_norm": 1042.4781494140625,
484
+ "learning_rate": 1.1736481776669307e-05,
485
+ "loss": 11.3515,
486
+ "step": 26
487
+ },
488
+ {
489
+ "epoch": 4.0,
490
+ "grad_norm": 663.6896362304688,
491
+ "learning_rate": 1.1045284632676535e-05,
492
+ "loss": 8.813,
493
+ "step": 27
494
+ },
495
+ {
496
+ "epoch": 4.090909090909091,
497
+ "grad_norm": 799.2377319335938,
498
+ "learning_rate": 1.0348994967025012e-05,
499
+ "loss": 6.9124,
500
+ "step": 28
501
+ },
502
+ {
503
+ "epoch": 4.090909090909091,
504
+ "eval_dim_1024_cosine_accuracy@1": 0.47619047619047616,
505
+ "eval_dim_1024_cosine_accuracy@10": 0.5714285714285714,
506
+ "eval_dim_1024_cosine_accuracy@3": 0.47619047619047616,
507
+ "eval_dim_1024_cosine_accuracy@5": 0.47619047619047616,
508
+ "eval_dim_1024_cosine_map@100": 0.6149109740313521,
509
+ "eval_dim_1024_cosine_mrr@10": 0.49092970521541945,
510
+ "eval_dim_1024_cosine_ndcg@10": 0.5159384546892658,
511
+ "eval_dim_1024_cosine_precision@1": 0.47619047619047616,
512
+ "eval_dim_1024_cosine_precision@10": 0.4,
513
+ "eval_dim_1024_cosine_precision@3": 0.4603174603174603,
514
+ "eval_dim_1024_cosine_precision@5": 0.419047619047619,
515
+ "eval_dim_1024_cosine_recall@1": 0.07822039072039072,
516
+ "eval_dim_1024_cosine_recall@10": 0.4449023199023199,
517
+ "eval_dim_1024_cosine_recall@3": 0.21085164835164832,
518
+ "eval_dim_1024_cosine_recall@5": 0.27602258852258854,
519
+ "eval_dim_128_cosine_accuracy@1": 0.47619047619047616,
520
+ "eval_dim_128_cosine_accuracy@10": 0.5238095238095238,
521
+ "eval_dim_128_cosine_accuracy@3": 0.47619047619047616,
522
+ "eval_dim_128_cosine_accuracy@5": 0.47619047619047616,
523
+ "eval_dim_128_cosine_map@100": 0.6025310247157158,
524
+ "eval_dim_128_cosine_mrr@10": 0.48299319727891155,
525
+ "eval_dim_128_cosine_ndcg@10": 0.49922091065744895,
526
+ "eval_dim_128_cosine_precision@1": 0.47619047619047616,
527
+ "eval_dim_128_cosine_precision@10": 0.3761904761904762,
528
+ "eval_dim_128_cosine_precision@3": 0.4603174603174603,
529
+ "eval_dim_128_cosine_precision@5": 0.419047619047619,
530
+ "eval_dim_128_cosine_recall@1": 0.07822039072039072,
531
+ "eval_dim_128_cosine_recall@10": 0.42506105006105005,
532
+ "eval_dim_128_cosine_recall@3": 0.21085164835164832,
533
+ "eval_dim_128_cosine_recall@5": 0.27602258852258854,
534
+ "eval_dim_256_cosine_accuracy@1": 0.5238095238095238,
535
+ "eval_dim_256_cosine_accuracy@10": 0.5714285714285714,
536
+ "eval_dim_256_cosine_accuracy@3": 0.5238095238095238,
537
+ "eval_dim_256_cosine_accuracy@5": 0.5238095238095238,
538
+ "eval_dim_256_cosine_map@100": 0.6411393184007045,
539
+ "eval_dim_256_cosine_mrr@10": 0.5306122448979591,
540
+ "eval_dim_256_cosine_ndcg@10": 0.5468399582764966,
541
+ "eval_dim_256_cosine_precision@1": 0.5238095238095238,
542
+ "eval_dim_256_cosine_precision@10": 0.4238095238095239,
543
+ "eval_dim_256_cosine_precision@3": 0.5079365079365079,
544
+ "eval_dim_256_cosine_precision@5": 0.4666666666666666,
545
+ "eval_dim_256_cosine_recall@1": 0.08005189255189255,
546
+ "eval_dim_256_cosine_recall@10": 0.4433760683760684,
547
+ "eval_dim_256_cosine_recall@3": 0.21634615384615385,
548
+ "eval_dim_256_cosine_recall@5": 0.28518009768009767,
549
+ "eval_dim_512_cosine_accuracy@1": 0.47619047619047616,
550
+ "eval_dim_512_cosine_accuracy@10": 0.5238095238095238,
551
+ "eval_dim_512_cosine_accuracy@3": 0.47619047619047616,
552
+ "eval_dim_512_cosine_accuracy@5": 0.47619047619047616,
553
+ "eval_dim_512_cosine_map@100": 0.5978106306698094,
554
+ "eval_dim_512_cosine_mrr@10": 0.48299319727891155,
555
+ "eval_dim_512_cosine_ndcg@10": 0.49922091065744895,
556
+ "eval_dim_512_cosine_precision@1": 0.47619047619047616,
557
+ "eval_dim_512_cosine_precision@10": 0.3761904761904762,
558
+ "eval_dim_512_cosine_precision@3": 0.4603174603174603,
559
+ "eval_dim_512_cosine_precision@5": 0.419047619047619,
560
+ "eval_dim_512_cosine_recall@1": 0.07822039072039072,
561
+ "eval_dim_512_cosine_recall@10": 0.42506105006105005,
562
+ "eval_dim_512_cosine_recall@3": 0.21085164835164832,
563
+ "eval_dim_512_cosine_recall@5": 0.27602258852258854,
564
+ "eval_dim_64_cosine_accuracy@1": 0.47619047619047616,
565
+ "eval_dim_64_cosine_accuracy@10": 0.5238095238095238,
566
+ "eval_dim_64_cosine_accuracy@3": 0.47619047619047616,
567
+ "eval_dim_64_cosine_accuracy@5": 0.47619047619047616,
568
+ "eval_dim_64_cosine_map@100": 0.5960251374266525,
569
+ "eval_dim_64_cosine_mrr@10": 0.48299319727891155,
570
+ "eval_dim_64_cosine_ndcg@10": 0.49922091065744895,
571
+ "eval_dim_64_cosine_precision@1": 0.47619047619047616,
572
+ "eval_dim_64_cosine_precision@10": 0.3761904761904762,
573
+ "eval_dim_64_cosine_precision@3": 0.4603174603174603,
574
+ "eval_dim_64_cosine_precision@5": 0.419047619047619,
575
+ "eval_dim_64_cosine_recall@1": 0.07822039072039072,
576
+ "eval_dim_64_cosine_recall@10": 0.42506105006105005,
577
+ "eval_dim_64_cosine_recall@3": 0.21085164835164832,
578
+ "eval_dim_64_cosine_recall@5": 0.27602258852258854,
579
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
580
+ "eval_dim_768_cosine_accuracy@10": 0.5714285714285714,
581
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
582
+ "eval_dim_768_cosine_accuracy@5": 0.5238095238095238,
583
+ "eval_dim_768_cosine_map@100": 0.6351788392177582,
584
+ "eval_dim_768_cosine_mrr@10": 0.5306122448979591,
585
+ "eval_dim_768_cosine_ndcg@10": 0.5468399582764966,
586
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
587
+ "eval_dim_768_cosine_precision@10": 0.4238095238095238,
588
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
589
+ "eval_dim_768_cosine_precision@5": 0.4666666666666666,
590
+ "eval_dim_768_cosine_recall@1": 0.08218864468864469,
591
+ "eval_dim_768_cosine_recall@10": 0.46474358974358976,
592
+ "eval_dim_768_cosine_recall@3": 0.22275641025641024,
593
+ "eval_dim_768_cosine_recall@5": 0.2958638583638584,
594
+ "eval_runtime": 7.0378,
595
+ "eval_samples_per_second": 0.0,
596
+ "eval_sequential_score": 0.49922091065744895,
597
+ "eval_steps_per_second": 0.0,
598
+ "step": 28
599
+ }
600
+ ],
601
+ "logging_steps": 1,
602
+ "max_steps": 50,
603
+ "num_input_tokens_seen": 0,
604
+ "num_train_epochs": 10,
605
+ "save_steps": 500,
606
+ "stateful_callbacks": {
607
+ "EarlyStoppingCallback": {
608
+ "args": {
609
+ "early_stopping_patience": 3,
610
+ "early_stopping_threshold": 0.0
611
+ },
612
+ "attributes": {
613
+ "early_stopping_patience_counter": 2
614
+ }
615
+ },
616
+ "TrainerControl": {
617
+ "args": {
618
+ "should_epoch_stop": false,
619
+ "should_evaluate": false,
620
+ "should_log": false,
621
+ "should_save": true,
622
+ "should_training_stop": false
623
+ },
624
+ "attributes": {}
625
+ }
626
+ },
627
+ "total_flos": 0.0,
628
+ "train_batch_size": 8,
629
+ "trial_name": null,
630
+ "trial_params": null
631
+ }
checkpoint-28/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14debc6c3f8c5edee5db8d97a3a78a007d313a13e4b96f43026da543b59bef8c
3
+ size 6097
checkpoint-35/1_Pooling/config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "word_embedding_dimension": 1024,
3
+ "pooling_mode_cls_token": false,
4
+ "pooling_mode_mean_tokens": true,
5
+ "pooling_mode_max_tokens": false,
6
+ "pooling_mode_mean_sqrt_len_tokens": false,
7
+ "pooling_mode_weightedmean_tokens": false,
8
+ "pooling_mode_lasttoken": false,
9
+ "include_prompt": true
10
+ }
checkpoint-35/README.md ADDED
@@ -0,0 +1,1658 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license: apache-2.0
5
+ tags:
6
+ - sentence-transformers
7
+ - sentence-similarity
8
+ - feature-extraction
9
+ - dense
10
+ - generated_from_trainer
11
+ - dataset_size:82
12
+ - loss:MatryoshkaLoss
13
+ - loss:MultipleNegativesRankingLoss
14
+ base_model: intfloat/multilingual-e5-large
15
+ widget:
16
+ - source_sentence: When did the victims give away credentials?
17
+ sentences:
18
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
19
+
20
+
21
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
22
+ benefit, causes damage to another’s property by persuading someone to act, omit,
23
+ or tolerate something through the knowing misrepresentation of false facts as
24
+ true, or through the unlawful concealment or suppression of true facts, shall
25
+ be punished by imprisonment of at least three months, and if the damage caused
26
+ is particularly large, by imprisonment of at least two years."
27
+
28
+
29
+ From this provision it follows that, for the crime of fraud to be established,
30
+ the following elements are required:
31
+
32
+
33
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
34
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
35
+
36
+
37
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
38
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
39
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
40
+ to themselves or another; and
41
+
42
+
43
+ c) Damage to another person’s property, as defined under civil law, which must
44
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
45
+ not required that the person deceived and the person who suffered the damage be
46
+ the same individual.
47
+
48
+
49
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
50
+ relating to the past or present, and not to those that will occur in the future,
51
+ such as mere promises or contractual obligations. However, when such promises
52
+ or obligations are accompanied by false assurances and representations of other
53
+ false facts referring to the present or the past, in such a manner as to create
54
+ the impression of future fulfillment based on a false present situation fabricated
55
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
56
+ the crime of fraud is established.
57
+
58
+
59
+ The term “property” refers to the totality of a person’s economic assets that
60
+ possess monetary value, while damage to property means its reduction—specifically,
61
+ the difference between the monetary value the property had before the disposition
62
+ caused by the fraudulent conduct and the value remaining after it. Property damage
63
+ exists even if the victim possesses an active claim for restitution.
64
+
65
+
66
+ The time of commission of the fraud is considered to be the moment when the perpetrator
67
+ acted and completed their fraudulent conduct, namely when they made the false
68
+ representations that deceived the victim or a third party. Any subsequent moment
69
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
70
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
71
+ - 'Voice phishing involves manipulating victims over the phone. Attackers pose as
72
+ bank officials or authorities and use intimidation to extract financial details.
73
+
74
+
75
+ Scenario:
76
+
77
+ - Victims are coerced into giving away PINs, passwords, or other credentials under
78
+ false pretenses of legal or financial emergencies.'
79
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
80
+
81
+
82
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
83
+ benefit, causes damage to another’s property by persuading someone to act, omit,
84
+ or tolerate something through the knowing misrepresentation of false facts as
85
+ true, or through the unlawful concealment or suppression of true facts, shall
86
+ be punished by imprisonment of at least three months, and if the damage caused
87
+ is particularly large, by imprisonment of at least two years."
88
+
89
+
90
+ From this provision, it follows that, for the crime of fraud to be established,
91
+ the following elements are required:
92
+
93
+
94
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
95
+ pecuniary benefit, without requiring that the benefit actually materialize;
96
+
97
+
98
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
99
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
100
+ is deceived and performs an act, omission, or acquiescence; and
101
+
102
+
103
+ c) Damage to another’s property, according to civil law, which must be causally
104
+ connected to the perpetrator’s deceptive acts or omissions. It is not required
105
+ that the deceived person and the person who suffered the loss be the same.
106
+
107
+
108
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
109
+ relating to the past or present, and not to those that will occur in the future,
110
+ such as mere promises or contractual obligations. However, when such promises
111
+ or obligations are accompanied by false assurances and representations of other
112
+ false facts relating to the present or the past, in such a way as to create the
113
+ impression of future fulfillment, based on a false present situation fabricated
114
+ by the perpetrator—who has already made the decision not to fulfill their obligation—then
115
+ the crime of fraud is established.
116
+
117
+
118
+ The term “property” denotes the totality of a person’s economic assets possessing
119
+ monetary value, while damage to property refers to its reduction—specifically,
120
+ the difference between the property’s monetary value before the disposition caused
121
+ by the fraudulent conduct and its value afterward. Property damage exists even
122
+ if the victim has an active claim for its restitution.
123
+
124
+
125
+ The time of commission of fraud is considered to be the moment when the perpetrator
126
+ acted and completed the deceptive conduct, that is, when they made the false representations
127
+ which deceived the victim or a third party. Any later time at which the victim’s
128
+ financial loss occurred—thus completing the fraud—or the time when the harmful
129
+ act or omission of the deceived person took place, is irrelevant.
130
+
131
+
132
+ The reference to multiple modes of commission of fraud (i.e., both the misrepresentation
133
+ of false facts and the concealment of true ones) may create ambiguity and contradiction,
134
+ unless it is made clear from the overall findings that the offense was committed
135
+ in one particular manner, and that the reference to the other merely serves to
136
+ define the intent (mens rea) of the perpetrator—specifically, that the representations
137
+ were false.
138
+
139
+
140
+ Furthermore, a conviction must contain the specific and well-reasoned justification
141
+ required by Articles 93 paragraph 3 of the Constitution and 139 of the Code of
142
+ Criminal Procedure. The absence of such reasoning constitutes grounds for cassation
143
+ (appeal) under Article 510 paragraph 1(d) of the Code of Criminal Procedure, when
144
+ the judgment does not set out, with clarity, completeness, and consistency, the
145
+ factual circumstances established by the evidence, upon which the court based
146
+ its findings regarding the objective and subjective elements of the offense, the
147
+ evidence supporting those findings, and the legal reasoning through which those
148
+ facts were subsumed under the applicable substantive criminal provision.
149
+
150
+
151
+ For the existence of such reasoning, the explanatory and operative parts of the
152
+ decision may complement each other, as they form a single, unified whole.
153
+
154
+
155
+ The existence of intent (dolus) does not generally need to be specially justified,
156
+ since it is inherent in the will to bring about the factual circumstances constituting
157
+ the objective elements of the offense, and it is presumed from their realization
158
+ in each particular case—unless the law requires additional elements for criminal
159
+ liability, such as the act being committed with knowledge of a specific circumstance
160
+ (direct intent) or with the pursuit of a further purpose, i.e., the achievement
161
+ of an additional result (offenses requiring a special subjective element).
162
+
163
+
164
+ Furthermore, under Article 510 paragraph 1(e) of the Code of Criminal Procedure,
165
+ a misapplication of substantive criminal law also constitutes grounds for cassation.
166
+ Such misapplication occurs when the trial court incorrectly applies the law to
167
+ the facts it has found to be true, or when the violation occurs indirectly, namely
168
+ when the reasoning of the judgment—comprising the combination of its factual and
169
+ operative parts and relating to the elements and identity of the offense—contains
170
+ ambiguities, contradictions, or logical gaps, rendering it impossible to verify,
171
+ on appeal, whether the law was applied correctly. In such cases, the judgment
172
+ lacks a lawful basis.'
173
+ - source_sentence: What must be the outcome of the deception in relation to property
174
+ damage?
175
+ sentences:
176
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
177
+
178
+
179
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
180
+ benefit, causes damage to another’s property by persuading someone to act, omit,
181
+ or tolerate something through the knowing misrepresentation of false facts as
182
+ true, or through the unlawful concealment or suppression of true facts, shall
183
+ be punished by imprisonment of at least three months, and if the damage caused
184
+ is particularly large, by imprisonment of at least two years."
185
+
186
+
187
+ From this provision, it follows that, for the crime of fraud to be established,
188
+ the following elements are required:
189
+
190
+
191
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
192
+ pecuniary benefit, without requiring that the benefit actually materialize;
193
+
194
+
195
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
196
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
197
+ is deceived and performs an act, omission, or acquiescence; and
198
+
199
+
200
+ c) Damage to another’s property, according to civil law, which must be causally
201
+ connected to the perpetrator’s deceptive acts or omissions. It is not required
202
+ that the deceived person and the person who suffered the loss be the same.
203
+
204
+
205
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
206
+ relating to the past or present, and not to those that will occur in the future,
207
+ such as mere promises or contractual obligations. However, when such promises
208
+ or obligations are accompanied by false assurances and representations of other
209
+ false facts relating to the present or the past, in such a way as to create the
210
+ impression of future fulfillment, based on a false present situation fabricated
211
+ by the perpetrator—who has already made the decision not to fulfill their obligation—then
212
+ the crime of fraud is established.
213
+
214
+
215
+ The term “property” denotes the totality of a person’s economic assets possessing
216
+ monetary value, while damage to property refers to its reduction—specifically,
217
+ the difference between the property’s monetary value before the disposition caused
218
+ by the fraudulent conduct and its value afterward. Property damage exists even
219
+ if the victim has an active claim for its restitution.
220
+
221
+
222
+ The time of commission of fraud is considered to be the moment when the perpetrator
223
+ acted and completed the deceptive conduct, that is, when they made the false representations
224
+ which deceived the victim or a third party. Any later time at which the victim’s
225
+ financial loss occurred—thus completing the fraud—or the time when the harmful
226
+ act or omission of the deceived person took place, is irrelevant.
227
+
228
+
229
+ The reference to multiple modes of commission of fraud (i.e., both the misrepresentation
230
+ of false facts and the concealment of true ones) may create ambiguity and contradiction,
231
+ unless it is made clear from the overall findings that the offense was committed
232
+ in one particular manner, and that the reference to the other merely serves to
233
+ define the intent (mens rea) of the perpetrator—specifically, that the representations
234
+ were false.
235
+
236
+
237
+ Furthermore, a conviction must contain the specific and well-reasoned justification
238
+ required by Articles 93 paragraph 3 of the Constitution and 139 of the Code of
239
+ Criminal Procedure. The absence of such reasoning constitutes grounds for cassation
240
+ (appeal) under Article 510 paragraph 1(d) of the Code of Criminal Procedure, when
241
+ the judgment does not set out, with clarity, completeness, and consistency, the
242
+ factual circumstances established by the evidence, upon which the court based
243
+ its findings regarding the objective and subjective elements of the offense, the
244
+ evidence supporting those findings, and the legal reasoning through which those
245
+ facts were subsumed under the applicable substantive criminal provision.
246
+
247
+
248
+ For the existence of such reasoning, the explanatory and operative parts of the
249
+ decision may complement each other, as they form a single, unified whole.
250
+
251
+
252
+ The existence of intent (dolus) does not generally need to be specially justified,
253
+ since it is inherent in the will to bring about the factual circumstances constituting
254
+ the objective elements of the offense, and it is presumed from their realization
255
+ in each particular case—unless the law requires additional elements for criminal
256
+ liability, such as the act being committed with knowledge of a specific circumstance
257
+ (direct intent) or with the pursuit of a further purpose, i.e., the achievement
258
+ of an additional result (offenses requiring a special subjective element).
259
+
260
+
261
+ Furthermore, under Article 510 paragraph 1(e) of the Code of Criminal Procedure,
262
+ a misapplication of substantive criminal law also constitutes grounds for cassation.
263
+ Such misapplication occurs when the trial court incorrectly applies the law to
264
+ the facts it has found to be true, or when the violation occurs indirectly, namely
265
+ when the reasoning of the judgment—comprising the combination of its factual and
266
+ operative parts and relating to the elements and identity of the offense—contains
267
+ ambiguities, contradictions, or logical gaps, rendering it impossible to verify,
268
+ on appeal, whether the law was applied correctly. In such cases, the judgment
269
+ lacks a lawful basis.'
270
+ - 'According to Article 386 paragraph 1 of the Greek Penal Code,
271
+
272
+
273
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
274
+ benefit, causes damage to another’s property by persuading someone to act, omit,
275
+ or tolerate something through the knowing misrepresentation of false facts as
276
+ true, or through the unlawful concealment or suppression of true facts, shall
277
+ be punished by imprisonment of at least three months, and if the damage caused
278
+ is particularly large, by imprisonment of at least two years."
279
+
280
+
281
+ From these provisions, it follows that, for the crime of fraud to be established,
282
+ the following elements are required:
283
+
284
+
285
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
286
+ pecuniary benefit;
287
+
288
+
289
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
290
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
291
+ is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves
292
+ or another; and
293
+
294
+
295
+ c) Damage to another’s property, as defined under civil law, which must be causally
296
+ connected to the perpetrator’s deceptive acts.
297
+
298
+
299
+ From the above provisions, it is deduced that the crime of fraud is established
300
+ both objectively and subjectively through the knowing misrepresentation of false
301
+ facts as true, or the unlawful concealment or suppression of true ones, by which
302
+ another person is deceived and, as a result, performs an act, omission, or acquiescence
303
+ involving a disposition of property that directly and necessarily causes financial
304
+ damage to the deceived person or another, with the intent that the perpetrator
305
+ or another gain an unlawful benefit. It is irrelevant whether this intended benefit
306
+ was ultimately achieved.
307
+
308
+
309
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
310
+ relating to the past or present, and not to those expected to occur in the future,
311
+ such as mere promises or contractual obligations. The false fact must have existed
312
+ in the past or must be a present circumstance at the time it is asserted, and
313
+ cannot relate to the future.
314
+
315
+
316
+ However, when future circumstances—that is, promises or contractual obligations—are
317
+ accompanied by false assurances and representations of other false facts referring
318
+ to the present or past, in such a way as to create the impression of future fulfillment,
319
+ based on a false present situation or supposed ability of the perpetrator, who
320
+ had already made the decision not to fulfill their obligation, then the crime
321
+ of fraud is established.'
322
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
323
+
324
+
325
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
326
+ benefit, causes damage to another person’s property by persuading someone to act,
327
+ omit, or tolerate something through the knowing misrepresentation of false facts
328
+ as true, or through the unlawful concealment or suppression of true facts, shall
329
+ be punished by imprisonment of at least three months, and if the damage caused
330
+ is particularly large, by imprisonment of at least two years."
331
+
332
+
333
+ From this provision, it follows that for the crime of fraud to be established,
334
+ the following elements are required:
335
+
336
+
337
+ a) Intent of the perpetrator to obtain for themselves or another an unlawful pecuniary
338
+ benefit, regardless of whether this benefit was actually realized;
339
+
340
+
341
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
342
+ or suppression of true facts, as a result of which, as a causal factor, someone
343
+ is deceived and acts in a way that is detrimental to themselves or another (by
344
+ an act, omission, or acquiescence); and
345
+
346
+
347
+ c) Damage to another’s property, in the sense recognized by civil law, which must
348
+ be causally linked to the fraudulent conduct (the deceptive act or omission of
349
+ the perpetrator) and to the resulting deception of the person who made the property
350
+ disposition. It is not required that the person deceived be the same person who
351
+ suffered the damage.
352
+
353
+
354
+ Property damage exists when there is a reduction or deterioration in the victim’s
355
+ assets, even if the victim has an active claim to restitution. However, as an
356
+ element of the objective aspect of the crime of fraud, the damage must be the
357
+ direct, necessary, and exclusive result of the property disposition—namely, the
358
+ act, omission, or acquiescence performed by the person deceived by the perpetrator’s
359
+ fraudulent conduct.
360
+
361
+
362
+ There must therefore be a causal connection between the perpetrator’s deceptive
363
+ behavior and the deception it caused, as well as between this deception and the
364
+ resulting property damage, which must be the direct, necessary, and exclusive
365
+ outcome of the deception and of the act, omission, or acquiescence of the deceived
366
+ person.
367
+
368
+
369
+ The term “facts” refers to real circumstances relating to the past or present,
370
+ and not to those expected to occur in the future, such as mere promises or contractual
371
+ obligations. However, when such promises or obligations are accompanied by false
372
+ assurances and representations of other false facts relating to the present or
373
+ the past, in such a way as to create the impression of future fulfillment, based
374
+ on the false present situation presented by a perpetrator who has already made
375
+ the decision not to fulfill their obligation, then the crime of fraud is established.
376
+
377
+
378
+ The time of commission of the fraud is considered to be the moment when the perpetrator
379
+ acted and completed their deceptive conduct—that is, when they made the false
380
+ representations that deceived the victim or a third party. Any later time at which
381
+ the victim’s financial loss actually occurred—thus completing the fraud—or the
382
+ time when the deceived person performed the harmful act or omission, is irrelevant.'
383
+ - source_sentence: How are victims tricked in email phishing scams?
384
+ sentences:
385
+ - 'According to Article 386 paragraph 1 of the Greek Penal Code,
386
+
387
+
388
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
389
+ benefit, causes damage to another’s property by persuading someone to act, omit,
390
+ or tolerate something through the knowing misrepresentation of false facts as
391
+ true, or through the unlawful concealment or suppression of true facts, shall
392
+ be punished by imprisonment of at least three months, and if the damage caused
393
+ is particularly large, by imprisonment of at least two years."
394
+
395
+
396
+ From these provisions, it follows that, for the crime of fraud to be established,
397
+ the following elements are required:
398
+
399
+
400
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
401
+ pecuniary benefit;
402
+
403
+
404
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
405
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
406
+ is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves
407
+ or another; and
408
+
409
+
410
+ c) Damage to another’s property, as defined under civil law, which must be causally
411
+ connected to the perpetrator’s deceptive acts.
412
+
413
+
414
+ From the above provisions, it is deduced that the crime of fraud is established
415
+ both objectively and subjectively through the knowing misrepresentation of false
416
+ facts as true, or the unlawful concealment or suppression of true ones, by which
417
+ another person is deceived and, as a result, performs an act, omission, or acquiescence
418
+ involving a disposition of property that directly and necessarily causes financial
419
+ damage to the deceived person or another, with the intent that the perpetrator
420
+ or another gain an unlawful benefit. It is irrelevant whether this intended benefit
421
+ was ultimately achieved.
422
+
423
+
424
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
425
+ relating to the past or present, and not to those expected to occur in the future,
426
+ such as mere promises or contractual obligations. The false fact must have existed
427
+ in the past or must be a present circumstance at the time it is asserted, and
428
+ cannot relate to the future.
429
+
430
+
431
+ However, when future circumstances—that is, promises or contractual obligations—are
432
+ accompanied by false assurances and representations of other false facts referring
433
+ to the present or past, in such a way as to create the impression of future fulfillment,
434
+ based on a false present situation or supposed ability of the perpetrator, who
435
+ had already made the decision not to fulfill their obligation, then the crime
436
+ of fraud is established.'
437
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
438
+
439
+
440
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
441
+ benefit, causes damage to another’s property by persuading someone to act, omit,
442
+ or tolerate something through the knowing misrepresentation of false facts as
443
+ true, or through the unlawful concealment or suppression of true facts, shall
444
+ be punished by imprisonment of at least three months, and if the damage caused
445
+ is particularly large, by imprisonment of at least two years."
446
+
447
+
448
+ From this provision it follows that, for the crime of fraud to be established,
449
+ the following elements are required:
450
+
451
+
452
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
453
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
454
+
455
+
456
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
457
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
458
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
459
+ to themselves or another; and
460
+
461
+
462
+ c) Damage to another person’s property, as defined under civil law, which must
463
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
464
+ not required that the person deceived and the person who suffered the damage be
465
+ the same individual.
466
+
467
+
468
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
469
+ relating to the past or present, and not to those that will occur in the future,
470
+ such as mere promises or contractual obligations. However, when such promises
471
+ or obligations are accompanied by false assurances and representations of other
472
+ false facts referring to the present or the past, in such a manner as to create
473
+ the impression of future fulfillment based on a false present situation fabricated
474
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
475
+ the crime of fraud is established.
476
+
477
+
478
+ The term “property” refers to the totality of a person’s economic assets that
479
+ possess monetary value, while damage to property means its reduction—specifically,
480
+ the difference between the monetary value the property had before the disposition
481
+ caused by the fraudulent conduct and the value remaining after it. Property damage
482
+ exists even if the victim possesses an active claim for restitution.
483
+
484
+
485
+ The time of commission of the fraud is considered to be the moment when the perpetrator
486
+ acted and completed their fraudulent conduct, namely when they made the false
487
+ representations that deceived the victim or a third party. Any subsequent moment
488
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
489
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
490
+ - 'Email phishing is a type of identity theft scam conducted via email or SMS. The
491
+ attacker uses social engineering tactics such as impersonating trusted entities
492
+ and inducing urgency. Victims are tricked into disclosing personal information
493
+ or downloading malware.
494
+
495
+
496
+ Scenarios:
497
+
498
+ - Scenario 1: Emails impersonating high-ranking executives accuse victims of crimes
499
+ to coerce them into revealing information or opening malware-laden attachments.
500
+
501
+ - Scenario 2: Emails/SMS from fake banks or authorities alert victims of data
502
+ breaches, directing them to spoofed websites to input credentials.
503
+
504
+ - Scenario 3: SMS messages deliver disguised malware apps that harvest sensitive
505
+ data.
506
+
507
+ - Scenario 4: SMS links lead to pharming sites that mimic trusted brands and steal
508
+ login data through fake pop-ups.'
509
+ - source_sentence: What circumstances do the term 'facts' refer to within the meaning
510
+ of the provision?
511
+ sentences:
512
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
513
+
514
+
515
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
516
+ benefit, causes damage to another person’s property by persuading someone to act,
517
+ omit, or tolerate something through the knowing misrepresentation of false facts
518
+ as true, or through the unlawful concealment or suppression of true facts, shall
519
+ be punished by imprisonment of at least three months, and if the damage caused
520
+ is particularly large, by imprisonment of at least two years."
521
+
522
+
523
+ From this provision, it follows that for the crime of fraud to be established,
524
+ the following elements are required:
525
+
526
+
527
+ a) Intent of the perpetrator to obtain for themselves or another an unlawful pecuniary
528
+ benefit, regardless of whether this benefit was actually realized;
529
+
530
+
531
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
532
+ or suppression of true facts, as a result of which, as a causal factor, someone
533
+ is deceived and acts in a way that is detrimental to themselves or another (by
534
+ an act, omission, or acquiescence); and
535
+
536
+
537
+ c) Damage to another’s property, in the sense recognized by civil law, which must
538
+ be causally linked to the fraudulent conduct (the deceptive act or omission of
539
+ the perpetrator) and to the resulting deception of the person who made the property
540
+ disposition. It is not required that the person deceived be the same person who
541
+ suffered the damage.
542
+
543
+
544
+ Property damage exists when there is a reduction or deterioration in the victim’s
545
+ assets, even if the victim has an active claim to restitution. However, as an
546
+ element of the objective aspect of the crime of fraud, the damage must be the
547
+ direct, necessary, and exclusive result of the property disposition—namely, the
548
+ act, omission, or acquiescence performed by the person deceived by the perpetrator’s
549
+ fraudulent conduct.
550
+
551
+
552
+ There must therefore be a causal connection between the perpetrator’s deceptive
553
+ behavior and the deception it caused, as well as between this deception and the
554
+ resulting property damage, which must be the direct, necessary, and exclusive
555
+ outcome of the deception and of the act, omission, or acquiescence of the deceived
556
+ person.
557
+
558
+
559
+ The term “facts” refers to real circumstances relating to the past or present,
560
+ and not to those expected to occur in the future, such as mere promises or contractual
561
+ obligations. However, when such promises or obligations are accompanied by false
562
+ assurances and representations of other false facts relating to the present or
563
+ the past, in such a way as to create the impression of future fulfillment, based
564
+ on the false present situation presented by a perpetrator who has already made
565
+ the decision not to fulfill their obligation, then the crime of fraud is established.
566
+
567
+
568
+ The time of commission of the fraud is considered to be the moment when the perpetrator
569
+ acted and completed their deceptive conduct—that is, when they made the false
570
+ representations that deceived the victim or a third party. Any later time at which
571
+ the victim’s financial loss actually occurred—thus completing the fraud—or the
572
+ time when the deceived person performed the harmful act or omission, is irrelevant.'
573
+ - '1. Anyone who, by knowingly presenting false facts as true or by unlawfully concealing
574
+ or withholding true facts, damages another person''s property by persuading someone
575
+ to act, omission, or tolerance with the aim of obtaining, for themselves or another,
576
+ an unlawful financial gain from the damage to that property shall be punished
577
+ with imprisonment, "and if the damage caused is particularly great, with imprisonment
578
+ of at least three (3) months and a fine." .
579
+
580
+ If the damage caused exceeds a total of one hundred and twenty thousand (120,000)
581
+ euros, imprisonment of up to ten (10) years and a fine shall be imposed.
582
+
583
+ 2. If the fraud is directed directly against the legal entity of the Greek State,
584
+ legal entities governed by public law, or local government organizations, and
585
+ the damage caused exceeds a total of one hundred and twenty thousand (120,000)
586
+ euros, a prison sentence of at least ten (10) years and a fine of up to one thousand
587
+ (1,000) daily units shall be imposed. This offense shall be time-barred after
588
+ twenty (20) years.
589
+
590
+ '
591
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
592
+
593
+
594
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
595
+ benefit, causes damage to another’s property by persuading someone to act, omit,
596
+ or tolerate something through the knowing misrepresentation of false facts as
597
+ true, or through the unlawful concealment or suppression of true facts, shall
598
+ be punished by imprisonment of at least three months, and if the damage caused
599
+ is particularly large, by imprisonment of at least two years."
600
+
601
+
602
+ From this provision it follows that, for the crime of fraud to be established,
603
+ the following elements are required:
604
+
605
+
606
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
607
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
608
+
609
+
610
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
611
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
612
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
613
+ to themselves or another; and
614
+
615
+
616
+ c) Damage to another person’s property, as defined under civil law, which must
617
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
618
+ not required that the person deceived and the person who suffered the damage be
619
+ the same individual.
620
+
621
+
622
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
623
+ relating to the past or present, and not to those that will occur in the future,
624
+ such as mere promises or contractual obligations. However, when such promises
625
+ or obligations are accompanied by false assurances and representations of other
626
+ false facts referring to the present or the past, in such a manner as to create
627
+ the impression of future fulfillment based on a false present situation fabricated
628
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
629
+ the crime of fraud is established.
630
+
631
+
632
+ The term “property” refers to the totality of a person’s economic assets that
633
+ possess monetary value, while damage to property means its reduction—specifically,
634
+ the difference between the monetary value the property had before the disposition
635
+ caused by the fraudulent conduct and the value remaining after it. Property damage
636
+ exists even if the victim possesses an active claim for restitution.
637
+
638
+
639
+ The time of commission of the fraud is considered to be the moment when the perpetrator
640
+ acted and completed their fraudulent conduct, namely when they made the false
641
+ representations that deceived the victim or a third party. Any subsequent moment
642
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
643
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
644
+ - source_sentence: When is the time of commission of the fraud considered?
645
+ sentences:
646
+ - 'Spear phishing targets specific individuals or employees within an organization
647
+ using personalized, deceptive emails. Unlike mass phishing, these emails are crafted
648
+ to seem familiar and urgent.
649
+
650
+
651
+ Scenarios:
652
+
653
+ - CEO Fraud: Attackers impersonate executives to extract financial or sensitive
654
+ data from employees.
655
+
656
+ - Whaling: High-ranking executives are targeted using tailored fraud emails that
657
+ press for immediate action without verification.'
658
+ - 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,
659
+
660
+
661
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
662
+ benefit, causes damage to another’s property by persuading someone to act, omit,
663
+ or tolerate something through the knowing misrepresentation of false facts as
664
+ true, or through the unlawful concealment or suppression of true facts, shall
665
+ be punished by imprisonment of at least three months, and if the damage caused
666
+ is particularly large, by imprisonment of at least two years."
667
+
668
+
669
+ From this provision it follows that, for the crime of fraud to be established,
670
+ the following elements are required:
671
+
672
+
673
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
674
+ pecuniary benefit, without it being necessary that the benefit actually materialize;
675
+
676
+
677
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
678
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
679
+ is deceived and proceeds to an act, omission, or acquiescence that is detrimental
680
+ to themselves or another; and
681
+
682
+
683
+ c) Damage to another person’s property, as defined under civil law, which must
684
+ be causally linked to the deceptive acts or omissions of the perpetrator. It is
685
+ not required that the person deceived and the person who suffered the damage be
686
+ the same individual.
687
+
688
+
689
+ The term “facts”, within the meaning of the above provision, refers to real circumstances
690
+ relating to the past or present, and not to those that will occur in the future,
691
+ such as mere promises or contractual obligations. However, when such promises
692
+ or obligations are accompanied by false assurances and representations of other
693
+ false facts referring to the present or the past, in such a manner as to create
694
+ the impression of future fulfillment based on a false present situation fabricated
695
+ by the perpetrator, who has already formed the decision not to fulfill their obligation,
696
+ the crime of fraud is established.
697
+
698
+
699
+ The term “property” refers to the totality of a person’s economic assets that
700
+ possess monetary value, while damage to property means its reduction—specifically,
701
+ the difference between the monetary value the property had before the disposition
702
+ caused by the fraudulent conduct and the value remaining after it. Property damage
703
+ exists even if the victim possesses an active claim for restitution.
704
+
705
+
706
+ The time of commission of the fraud is considered to be the moment when the perpetrator
707
+ acted and completed their fraudulent conduct, namely when they made the false
708
+ representations that deceived the victim or a third party. Any subsequent moment
709
+ at which the victim’s damage actually occurred—thereby completing the fraud—or
710
+ the time when the victim carried out the harmful act or omission, is irrelevant.'
711
+ - 'According to Article 386 paragraph 1 of the Greek Penal Code,
712
+
713
+
714
+ "Whoever, with the intent to obtain for themselves or another an unlawful pecuniary
715
+ benefit, causes damage to another’s property by persuading someone to act, omit,
716
+ or tolerate something through the knowing misrepresentation of false facts as
717
+ true, or through the unlawful concealment or suppression of true facts, shall
718
+ be punished by imprisonment of at least three months, and if the damage caused
719
+ is particularly large, by imprisonment of at least two years."
720
+
721
+
722
+ From these provisions, it follows that, for the crime of fraud to be established,
723
+ the following elements are required:
724
+
725
+
726
+ a) The intent of the perpetrator to obtain for themselves or another an unlawful
727
+ pecuniary benefit;
728
+
729
+
730
+ b) The knowing misrepresentation of false facts as true, or the unlawful concealment
731
+ or suppression of true facts, as a result of which—serving as the causal factor—someone
732
+ is deceived and proceeds to an act, omission, or acquiescence detrimental to themselves
733
+ or another; and
734
+
735
+
736
+ c) Damage to another’s property, as defined under civil law, which must be causally
737
+ connected to the perpetrator’s deceptive acts.
738
+
739
+
740
+ From the above provisions, it is deduced that the crime of fraud is established
741
+ both objectively and subjectively through the knowing misrepresentation of false
742
+ facts as true, or the unlawful concealment or suppression of true ones, by which
743
+ another person is deceived and, as a result, performs an act, omission, or acquiescence
744
+ involving a disposition of property that directly and necessarily causes financial
745
+ damage to the deceived person or another, with the intent that the perpetrator
746
+ or another gain an unlawful benefit. It is irrelevant whether this intended benefit
747
+ was ultimately achieved.
748
+
749
+
750
+ The term “facts,” within the meaning of the above provision, refers to real circumstances
751
+ relating to the past or present, and not to those expected to occur in the future,
752
+ such as mere promises or contractual obligations. The false fact must have existed
753
+ in the past or must be a present circumstance at the time it is asserted, and
754
+ cannot relate to the future.
755
+
756
+
757
+ However, when future circumstances—that is, promises or contractual obligations—are
758
+ accompanied by false assurances and representations of other false facts referring
759
+ to the present or past, in such a way as to create the impression of future fulfillment,
760
+ based on a false present situation or supposed ability of the perpetrator, who
761
+ had already made the decision not to fulfill their obligation, then the crime
762
+ of fraud is established.'
763
+ pipeline_tag: sentence-similarity
764
+ library_name: sentence-transformers
765
+ metrics:
766
+ - cosine_accuracy@1
767
+ - cosine_accuracy@3
768
+ - cosine_accuracy@5
769
+ - cosine_accuracy@10
770
+ - cosine_precision@1
771
+ - cosine_precision@3
772
+ - cosine_precision@5
773
+ - cosine_precision@10
774
+ - cosine_recall@1
775
+ - cosine_recall@3
776
+ - cosine_recall@5
777
+ - cosine_recall@10
778
+ - cosine_ndcg@10
779
+ - cosine_mrr@10
780
+ - cosine_map@100
781
+ model-index:
782
+ - name: multilingual_e5_large Finetuned on Data
783
+ results:
784
+ - task:
785
+ type: information-retrieval
786
+ name: Information Retrieval
787
+ dataset:
788
+ name: dim 1024
789
+ type: dim_1024
790
+ metrics:
791
+ - type: cosine_accuracy@1
792
+ value: 0.5238095238095238
793
+ name: Cosine Accuracy@1
794
+ - type: cosine_accuracy@3
795
+ value: 0.5238095238095238
796
+ name: Cosine Accuracy@3
797
+ - type: cosine_accuracy@5
798
+ value: 0.5238095238095238
799
+ name: Cosine Accuracy@5
800
+ - type: cosine_accuracy@10
801
+ value: 0.6190476190476191
802
+ name: Cosine Accuracy@10
803
+ - type: cosine_precision@1
804
+ value: 0.5238095238095238
805
+ name: Cosine Precision@1
806
+ - type: cosine_precision@3
807
+ value: 0.5079365079365079
808
+ name: Cosine Precision@3
809
+ - type: cosine_precision@5
810
+ value: 0.4666666666666666
811
+ name: Cosine Precision@5
812
+ - type: cosine_precision@10
813
+ value: 0.4428571428571429
814
+ name: Cosine Precision@10
815
+ - type: cosine_recall@1
816
+ value: 0.08218864468864469
817
+ name: Cosine Recall@1
818
+ - type: cosine_recall@3
819
+ value: 0.22275641025641024
820
+ name: Cosine Recall@3
821
+ - type: cosine_recall@5
822
+ value: 0.2958638583638584
823
+ name: Cosine Recall@5
824
+ - type: cosine_recall@10
825
+ value: 0.4766483516483517
826
+ name: Cosine Recall@10
827
+ - type: cosine_ndcg@10
828
+ value: 0.5598242514045669
829
+ name: Cosine Ndcg@10
830
+ - type: cosine_mrr@10
831
+ value: 0.5374149659863945
832
+ name: Cosine Mrr@10
833
+ - type: cosine_map@100
834
+ value: 0.6534286699882501
835
+ name: Cosine Map@100
836
+ - task:
837
+ type: information-retrieval
838
+ name: Information Retrieval
839
+ dataset:
840
+ name: dim 768
841
+ type: dim_768
842
+ metrics:
843
+ - type: cosine_accuracy@1
844
+ value: 0.5238095238095238
845
+ name: Cosine Accuracy@1
846
+ - type: cosine_accuracy@3
847
+ value: 0.5238095238095238
848
+ name: Cosine Accuracy@3
849
+ - type: cosine_accuracy@5
850
+ value: 0.5238095238095238
851
+ name: Cosine Accuracy@5
852
+ - type: cosine_accuracy@10
853
+ value: 0.6190476190476191
854
+ name: Cosine Accuracy@10
855
+ - type: cosine_precision@1
856
+ value: 0.5238095238095238
857
+ name: Cosine Precision@1
858
+ - type: cosine_precision@3
859
+ value: 0.5079365079365079
860
+ name: Cosine Precision@3
861
+ - type: cosine_precision@5
862
+ value: 0.4666666666666666
863
+ name: Cosine Precision@5
864
+ - type: cosine_precision@10
865
+ value: 0.4428571428571429
866
+ name: Cosine Precision@10
867
+ - type: cosine_recall@1
868
+ value: 0.08218864468864469
869
+ name: Cosine Recall@1
870
+ - type: cosine_recall@3
871
+ value: 0.22275641025641024
872
+ name: Cosine Recall@3
873
+ - type: cosine_recall@5
874
+ value: 0.2958638583638584
875
+ name: Cosine Recall@5
876
+ - type: cosine_recall@10
877
+ value: 0.4766483516483517
878
+ name: Cosine Recall@10
879
+ - type: cosine_ndcg@10
880
+ value: 0.5598242514045669
881
+ name: Cosine Ndcg@10
882
+ - type: cosine_mrr@10
883
+ value: 0.5374149659863945
884
+ name: Cosine Mrr@10
885
+ - type: cosine_map@100
886
+ value: 0.653075337994289
887
+ name: Cosine Map@100
888
+ - task:
889
+ type: information-retrieval
890
+ name: Information Retrieval
891
+ dataset:
892
+ name: dim 512
893
+ type: dim_512
894
+ metrics:
895
+ - type: cosine_accuracy@1
896
+ value: 0.5238095238095238
897
+ name: Cosine Accuracy@1
898
+ - type: cosine_accuracy@3
899
+ value: 0.5238095238095238
900
+ name: Cosine Accuracy@3
901
+ - type: cosine_accuracy@5
902
+ value: 0.5238095238095238
903
+ name: Cosine Accuracy@5
904
+ - type: cosine_accuracy@10
905
+ value: 0.6190476190476191
906
+ name: Cosine Accuracy@10
907
+ - type: cosine_precision@1
908
+ value: 0.5238095238095238
909
+ name: Cosine Precision@1
910
+ - type: cosine_precision@3
911
+ value: 0.5079365079365079
912
+ name: Cosine Precision@3
913
+ - type: cosine_precision@5
914
+ value: 0.4666666666666666
915
+ name: Cosine Precision@5
916
+ - type: cosine_precision@10
917
+ value: 0.4428571428571429
918
+ name: Cosine Precision@10
919
+ - type: cosine_recall@1
920
+ value: 0.08218864468864469
921
+ name: Cosine Recall@1
922
+ - type: cosine_recall@3
923
+ value: 0.22275641025641024
924
+ name: Cosine Recall@3
925
+ - type: cosine_recall@5
926
+ value: 0.2958638583638584
927
+ name: Cosine Recall@5
928
+ - type: cosine_recall@10
929
+ value: 0.4766483516483517
930
+ name: Cosine Recall@10
931
+ - type: cosine_ndcg@10
932
+ value: 0.5598242514045669
933
+ name: Cosine Ndcg@10
934
+ - type: cosine_mrr@10
935
+ value: 0.5374149659863945
936
+ name: Cosine Mrr@10
937
+ - type: cosine_map@100
938
+ value: 0.6492208787775379
939
+ name: Cosine Map@100
940
+ - task:
941
+ type: information-retrieval
942
+ name: Information Retrieval
943
+ dataset:
944
+ name: dim 256
945
+ type: dim_256
946
+ metrics:
947
+ - type: cosine_accuracy@1
948
+ value: 0.6190476190476191
949
+ name: Cosine Accuracy@1
950
+ - type: cosine_accuracy@3
951
+ value: 0.6190476190476191
952
+ name: Cosine Accuracy@3
953
+ - type: cosine_accuracy@5
954
+ value: 0.6190476190476191
955
+ name: Cosine Accuracy@5
956
+ - type: cosine_accuracy@10
957
+ value: 0.6666666666666666
958
+ name: Cosine Accuracy@10
959
+ - type: cosine_precision@1
960
+ value: 0.6190476190476191
961
+ name: Cosine Precision@1
962
+ - type: cosine_precision@3
963
+ value: 0.6031746031746031
964
+ name: Cosine Precision@3
965
+ - type: cosine_precision@5
966
+ value: 0.5619047619047619
967
+ name: Cosine Precision@5
968
+ - type: cosine_precision@10
969
+ value: 0.5190476190476192
970
+ name: Cosine Precision@10
971
+ - type: cosine_recall@1
972
+ value: 0.08600427350427349
973
+ name: Cosine Recall@1
974
+ - type: cosine_recall@3
975
+ value: 0.2342032967032967
976
+ name: Cosine Recall@3
977
+ - type: cosine_recall@5
978
+ value: 0.31494200244200243
979
+ name: Cosine Recall@5
980
+ - type: cosine_recall@10
981
+ value: 0.5028998778998779
982
+ name: Cosine Recall@10
983
+ - type: cosine_ndcg@10
984
+ value: 0.6420780535145918
985
+ name: Cosine Ndcg@10
986
+ - type: cosine_mrr@10
987
+ value: 0.6258503401360545
988
+ name: Cosine Mrr@10
989
+ - type: cosine_map@100
990
+ value: 0.6975707466438095
991
+ name: Cosine Map@100
992
+ - task:
993
+ type: information-retrieval
994
+ name: Information Retrieval
995
+ dataset:
996
+ name: dim 128
997
+ type: dim_128
998
+ metrics:
999
+ - type: cosine_accuracy@1
1000
+ value: 0.5238095238095238
1001
+ name: Cosine Accuracy@1
1002
+ - type: cosine_accuracy@3
1003
+ value: 0.5238095238095238
1004
+ name: Cosine Accuracy@3
1005
+ - type: cosine_accuracy@5
1006
+ value: 0.5238095238095238
1007
+ name: Cosine Accuracy@5
1008
+ - type: cosine_accuracy@10
1009
+ value: 0.6190476190476191
1010
+ name: Cosine Accuracy@10
1011
+ - type: cosine_precision@1
1012
+ value: 0.5238095238095238
1013
+ name: Cosine Precision@1
1014
+ - type: cosine_precision@3
1015
+ value: 0.5079365079365079
1016
+ name: Cosine Precision@3
1017
+ - type: cosine_precision@5
1018
+ value: 0.4666666666666666
1019
+ name: Cosine Precision@5
1020
+ - type: cosine_precision@10
1021
+ value: 0.4428571428571429
1022
+ name: Cosine Precision@10
1023
+ - type: cosine_recall@1
1024
+ value: 0.0811965811965812
1025
+ name: Cosine Recall@1
1026
+ - type: cosine_recall@3
1027
+ value: 0.21978021978021975
1028
+ name: Cosine Recall@3
1029
+ - type: cosine_recall@5
1030
+ value: 0.2909035409035409
1031
+ name: Cosine Recall@5
1032
+ - type: cosine_recall@10
1033
+ value: 0.46672771672771673
1034
+ name: Cosine Recall@10
1035
+ - type: cosine_ndcg@10
1036
+ value: 0.5598242514045669
1037
+ name: Cosine Ndcg@10
1038
+ - type: cosine_mrr@10
1039
+ value: 0.5374149659863945
1040
+ name: Cosine Mrr@10
1041
+ - type: cosine_map@100
1042
+ value: 0.6478872365910466
1043
+ name: Cosine Map@100
1044
+ - task:
1045
+ type: information-retrieval
1046
+ name: Information Retrieval
1047
+ dataset:
1048
+ name: dim 64
1049
+ type: dim_64
1050
+ metrics:
1051
+ - type: cosine_accuracy@1
1052
+ value: 0.42857142857142855
1053
+ name: Cosine Accuracy@1
1054
+ - type: cosine_accuracy@3
1055
+ value: 0.47619047619047616
1056
+ name: Cosine Accuracy@3
1057
+ - type: cosine_accuracy@5
1058
+ value: 0.47619047619047616
1059
+ name: Cosine Accuracy@5
1060
+ - type: cosine_accuracy@10
1061
+ value: 0.5714285714285714
1062
+ name: Cosine Accuracy@10
1063
+ - type: cosine_precision@1
1064
+ value: 0.42857142857142855
1065
+ name: Cosine Precision@1
1066
+ - type: cosine_precision@3
1067
+ value: 0.4444444444444445
1068
+ name: Cosine Precision@3
1069
+ - type: cosine_precision@5
1070
+ value: 0.419047619047619
1071
+ name: Cosine Precision@5
1072
+ - type: cosine_precision@10
1073
+ value: 0.3952380952380953
1074
+ name: Cosine Precision@10
1075
+ - type: cosine_recall@1
1076
+ value: 0.054410866910866905
1077
+ name: Cosine Recall@1
1078
+ - type: cosine_recall@3
1079
+ value: 0.18704212454212454
1080
+ name: Cosine Recall@3
1081
+ - type: cosine_recall@5
1082
+ value: 0.27602258852258854
1083
+ name: Cosine Recall@5
1084
+ - type: cosine_recall@10
1085
+ value: 0.43696581196581197
1086
+ name: Cosine Recall@10
1087
+ - type: cosine_ndcg@10
1088
+ value: 0.4917595713548203
1089
+ name: Cosine Ndcg@10
1090
+ - type: cosine_mrr@10
1091
+ value: 0.45804988662131524
1092
+ name: Cosine Mrr@10
1093
+ - type: cosine_map@100
1094
+ value: 0.5872011588310861
1095
+ name: Cosine Map@100
1096
+ ---
1097
+
1098
+ # multilingual_e5_large Finetuned on Data
1099
+
1100
+ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large). 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.
1101
+
1102
+ ## Model Details
1103
+
1104
+ ### Model Description
1105
+ - **Model Type:** Sentence Transformer
1106
+ - **Base model:** [intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large) <!-- at revision 0dc5580a448e4284468b8909bae50fa925907bc5 -->
1107
+ - **Maximum Sequence Length:** 512 tokens
1108
+ - **Output Dimensionality:** 1024 dimensions
1109
+ - **Similarity Function:** Cosine Similarity
1110
+ <!-- - **Training Dataset:** Unknown -->
1111
+ - **Language:** en
1112
+ - **License:** apache-2.0
1113
+
1114
+ ### Model Sources
1115
+
1116
+ - **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
1117
+ - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
1118
+ - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
1119
+
1120
+ ### Full Model Architecture
1121
+
1122
+ ```
1123
+ SentenceTransformer(
1124
+ (0): Transformer({'max_seq_length': 512, 'do_lower_case': False, 'architecture': 'XLMRobertaModel'})
1125
+ (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
1126
+ (2): Normalize()
1127
+ )
1128
+ ```
1129
+
1130
+ ## Usage
1131
+
1132
+ ### Direct Usage (Sentence Transformers)
1133
+
1134
+ First install the Sentence Transformers library:
1135
+
1136
+ ```bash
1137
+ pip install -U sentence-transformers
1138
+ ```
1139
+
1140
+ Then you can load this model and run inference.
1141
+ ```python
1142
+ from sentence_transformers import SentenceTransformer
1143
+
1144
+ # Download from the 🤗 Hub
1145
+ model = SentenceTransformer("sentence_transformers_model_id")
1146
+ # Run inference
1147
+ sentences = [
1148
+ 'When is the time of commission of the fraud considered?',
1149
+ 'According to the provision of Article 386 paragraph 1 of the Greek Penal Code,\n\n"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."\n\nFrom this provision it follows that, for the crime of fraud to be established, the following elements are required:\n\na) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit, without it being necessary that the benefit actually materialize;\n\nb) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence that is detrimental to themselves or another; and\n\nc) Damage to another person’s property, as defined under civil law, which must be causally linked to the deceptive acts or omissions of the perpetrator. It is not required that the person deceived and the person who suffered the damage be the same individual.\n\nThe term “facts”, within the meaning of the above provision, refers to real circumstances relating to the past or present, and not to those that will occur in the future, such as mere promises or contractual obligations. However, when such promises or obligations are accompanied by false assurances and representations of other false facts referring to the present or the past, in such a manner as to create the impression of future fulfillment based on a false present situation fabricated by the perpetrator, who has already formed the decision not to fulfill their obligation, the crime of fraud is established.\n\nThe term “property” refers to the totality of a person’s economic assets that possess monetary value, while damage to property means its reduction—specifically, the difference between the monetary value the property had before the disposition caused by the fraudulent conduct and the value remaining after it. Property damage exists even if the victim possesses an active claim for restitution.\n\nThe time of commission of the fraud is considered to be the moment when the perpetrator acted and completed their fraudulent conduct, namely when they made the false representations that deceived the victim or a third party. Any subsequent moment at which the victim’s damage actually occurred—thereby completing the fraud—or the time when the victim carried out the harmful act or omission, is irrelevant.',
1150
+ 'Spear phishing targets specific individuals or employees within an organization using personalized, deceptive emails. Unlike mass phishing, these emails are crafted to seem familiar and urgent.\n\nScenarios:\n- CEO Fraud: Attackers impersonate executives to extract financial or sensitive data from employees.\n- Whaling: High-ranking executives are targeted using tailored fraud emails that press for immediate action without verification.',
1151
+ ]
1152
+ embeddings = model.encode(sentences)
1153
+ print(embeddings.shape)
1154
+ # [3, 1024]
1155
+
1156
+ # Get the similarity scores for the embeddings
1157
+ similarities = model.similarity(embeddings, embeddings)
1158
+ print(similarities)
1159
+ # tensor([[1.0000, 0.5608, 0.2769],
1160
+ # [0.5608, 1.0000, 0.3160],
1161
+ # [0.2769, 0.3160, 1.0001]])
1162
+ ```
1163
+
1164
+ <!--
1165
+ ### Direct Usage (Transformers)
1166
+
1167
+ <details><summary>Click to see the direct usage in Transformers</summary>
1168
+
1169
+ </details>
1170
+ -->
1171
+
1172
+ <!--
1173
+ ### Downstream Usage (Sentence Transformers)
1174
+
1175
+ You can finetune this model on your own dataset.
1176
+
1177
+ <details><summary>Click to expand</summary>
1178
+
1179
+ </details>
1180
+ -->
1181
+
1182
+ <!--
1183
+ ### Out-of-Scope Use
1184
+
1185
+ *List how the model may foreseeably be misused and address what users ought not to do with the model.*
1186
+ -->
1187
+
1188
+ ## Evaluation
1189
+
1190
+ ### Metrics
1191
+
1192
+ #### Information Retrieval
1193
+
1194
+ * Dataset: `dim_1024`
1195
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1196
+ ```json
1197
+ {
1198
+ "truncate_dim": 1024
1199
+ }
1200
+ ```
1201
+
1202
+ | Metric | Value |
1203
+ |:--------------------|:-----------|
1204
+ | cosine_accuracy@1 | 0.5238 |
1205
+ | cosine_accuracy@3 | 0.5238 |
1206
+ | cosine_accuracy@5 | 0.5238 |
1207
+ | cosine_accuracy@10 | 0.619 |
1208
+ | cosine_precision@1 | 0.5238 |
1209
+ | cosine_precision@3 | 0.5079 |
1210
+ | cosine_precision@5 | 0.4667 |
1211
+ | cosine_precision@10 | 0.4429 |
1212
+ | cosine_recall@1 | 0.0822 |
1213
+ | cosine_recall@3 | 0.2228 |
1214
+ | cosine_recall@5 | 0.2959 |
1215
+ | cosine_recall@10 | 0.4766 |
1216
+ | **cosine_ndcg@10** | **0.5598** |
1217
+ | cosine_mrr@10 | 0.5374 |
1218
+ | cosine_map@100 | 0.6534 |
1219
+
1220
+ #### Information Retrieval
1221
+
1222
+ * Dataset: `dim_768`
1223
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1224
+ ```json
1225
+ {
1226
+ "truncate_dim": 768
1227
+ }
1228
+ ```
1229
+
1230
+ | Metric | Value |
1231
+ |:--------------------|:-----------|
1232
+ | cosine_accuracy@1 | 0.5238 |
1233
+ | cosine_accuracy@3 | 0.5238 |
1234
+ | cosine_accuracy@5 | 0.5238 |
1235
+ | cosine_accuracy@10 | 0.619 |
1236
+ | cosine_precision@1 | 0.5238 |
1237
+ | cosine_precision@3 | 0.5079 |
1238
+ | cosine_precision@5 | 0.4667 |
1239
+ | cosine_precision@10 | 0.4429 |
1240
+ | cosine_recall@1 | 0.0822 |
1241
+ | cosine_recall@3 | 0.2228 |
1242
+ | cosine_recall@5 | 0.2959 |
1243
+ | cosine_recall@10 | 0.4766 |
1244
+ | **cosine_ndcg@10** | **0.5598** |
1245
+ | cosine_mrr@10 | 0.5374 |
1246
+ | cosine_map@100 | 0.6531 |
1247
+
1248
+ #### Information Retrieval
1249
+
1250
+ * Dataset: `dim_512`
1251
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1252
+ ```json
1253
+ {
1254
+ "truncate_dim": 512
1255
+ }
1256
+ ```
1257
+
1258
+ | Metric | Value |
1259
+ |:--------------------|:-----------|
1260
+ | cosine_accuracy@1 | 0.5238 |
1261
+ | cosine_accuracy@3 | 0.5238 |
1262
+ | cosine_accuracy@5 | 0.5238 |
1263
+ | cosine_accuracy@10 | 0.619 |
1264
+ | cosine_precision@1 | 0.5238 |
1265
+ | cosine_precision@3 | 0.5079 |
1266
+ | cosine_precision@5 | 0.4667 |
1267
+ | cosine_precision@10 | 0.4429 |
1268
+ | cosine_recall@1 | 0.0822 |
1269
+ | cosine_recall@3 | 0.2228 |
1270
+ | cosine_recall@5 | 0.2959 |
1271
+ | cosine_recall@10 | 0.4766 |
1272
+ | **cosine_ndcg@10** | **0.5598** |
1273
+ | cosine_mrr@10 | 0.5374 |
1274
+ | cosine_map@100 | 0.6492 |
1275
+
1276
+ #### Information Retrieval
1277
+
1278
+ * Dataset: `dim_256`
1279
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1280
+ ```json
1281
+ {
1282
+ "truncate_dim": 256
1283
+ }
1284
+ ```
1285
+
1286
+ | Metric | Value |
1287
+ |:--------------------|:-----------|
1288
+ | cosine_accuracy@1 | 0.619 |
1289
+ | cosine_accuracy@3 | 0.619 |
1290
+ | cosine_accuracy@5 | 0.619 |
1291
+ | cosine_accuracy@10 | 0.6667 |
1292
+ | cosine_precision@1 | 0.619 |
1293
+ | cosine_precision@3 | 0.6032 |
1294
+ | cosine_precision@5 | 0.5619 |
1295
+ | cosine_precision@10 | 0.519 |
1296
+ | cosine_recall@1 | 0.086 |
1297
+ | cosine_recall@3 | 0.2342 |
1298
+ | cosine_recall@5 | 0.3149 |
1299
+ | cosine_recall@10 | 0.5029 |
1300
+ | **cosine_ndcg@10** | **0.6421** |
1301
+ | cosine_mrr@10 | 0.6259 |
1302
+ | cosine_map@100 | 0.6976 |
1303
+
1304
+ #### Information Retrieval
1305
+
1306
+ * Dataset: `dim_128`
1307
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1308
+ ```json
1309
+ {
1310
+ "truncate_dim": 128
1311
+ }
1312
+ ```
1313
+
1314
+ | Metric | Value |
1315
+ |:--------------------|:-----------|
1316
+ | cosine_accuracy@1 | 0.5238 |
1317
+ | cosine_accuracy@3 | 0.5238 |
1318
+ | cosine_accuracy@5 | 0.5238 |
1319
+ | cosine_accuracy@10 | 0.619 |
1320
+ | cosine_precision@1 | 0.5238 |
1321
+ | cosine_precision@3 | 0.5079 |
1322
+ | cosine_precision@5 | 0.4667 |
1323
+ | cosine_precision@10 | 0.4429 |
1324
+ | cosine_recall@1 | 0.0812 |
1325
+ | cosine_recall@3 | 0.2198 |
1326
+ | cosine_recall@5 | 0.2909 |
1327
+ | cosine_recall@10 | 0.4667 |
1328
+ | **cosine_ndcg@10** | **0.5598** |
1329
+ | cosine_mrr@10 | 0.5374 |
1330
+ | cosine_map@100 | 0.6479 |
1331
+
1332
+ #### Information Retrieval
1333
+
1334
+ * Dataset: `dim_64`
1335
+ * Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator) with these parameters:
1336
+ ```json
1337
+ {
1338
+ "truncate_dim": 64
1339
+ }
1340
+ ```
1341
+
1342
+ | Metric | Value |
1343
+ |:--------------------|:-----------|
1344
+ | cosine_accuracy@1 | 0.4286 |
1345
+ | cosine_accuracy@3 | 0.4762 |
1346
+ | cosine_accuracy@5 | 0.4762 |
1347
+ | cosine_accuracy@10 | 0.5714 |
1348
+ | cosine_precision@1 | 0.4286 |
1349
+ | cosine_precision@3 | 0.4444 |
1350
+ | cosine_precision@5 | 0.419 |
1351
+ | cosine_precision@10 | 0.3952 |
1352
+ | cosine_recall@1 | 0.0544 |
1353
+ | cosine_recall@3 | 0.187 |
1354
+ | cosine_recall@5 | 0.276 |
1355
+ | cosine_recall@10 | 0.437 |
1356
+ | **cosine_ndcg@10** | **0.4918** |
1357
+ | cosine_mrr@10 | 0.458 |
1358
+ | cosine_map@100 | 0.5872 |
1359
+
1360
+ <!--
1361
+ ## Bias, Risks and Limitations
1362
+
1363
+ *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
1364
+ -->
1365
+
1366
+ <!--
1367
+ ### Recommendations
1368
+
1369
+ *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
1370
+ -->
1371
+
1372
+ ## Training Details
1373
+
1374
+ ### Training Dataset
1375
+
1376
+ #### Unnamed Dataset
1377
+
1378
+ * Size: 82 training samples
1379
+ * Columns: <code>anchor</code> and <code>positive</code>
1380
+ * Approximate statistics based on the first 82 samples:
1381
+ | | anchor | positive |
1382
+ |:--------|:----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|
1383
+ | type | string | string |
1384
+ | details | <ul><li>min: 9 tokens</li><li>mean: 18.17 tokens</li><li>max: 34 tokens</li></ul> | <ul><li>min: 69 tokens</li><li>mean: 399.51 tokens</li><li>max: 512 tokens</li></ul> |
1385
+ * Samples:
1386
+ | anchor | positive |
1387
+ |:----------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
1388
+ | <code>What determines whether the act in question shall be punished if the offender is in the service of the legal holder of the data?</code> | <code>Everyone who obtains access to data recorded in a computer or in the external memory of a computer or transmitted by telecommunication systems shall be punished with imprisonment for up to six months or by a fine from 29 to 15,000 Euro, under the condition that these acts have been committed without right, especially in violation of prohibitions or of security measures taken by the legal holder. If the act concerns the international relations or the security of the State, he shall be punished according to Article 148.<br>If the offender is in the service of the legal holder of the data, the act of the preceding paragraph shall be punished only if it has been explicitly prohibited by internal regulations or by a written decision of the holder or of a competent employee of his.<br></code> |
1389
+ | <code>What must be causally connected to the perpetrator's deceptive acts?</code> | <code>According to Article 386 paragraph 1 of the Greek Penal Code,<br><br>"Whoever, with the intent to obtain for themselves or another an unlawful pecuniary benefit, causes damage to another’s property by persuading someone to act, omit, or tolerate something through the knowing misrepresentation of false facts as true, or through the unlawful concealment or suppression of true facts, shall be punished by imprisonment of at least three months, and if the damage caused is particularly large, by imprisonment of at least two years."<br><br>From these provisions, it follows that, for the crime of fraud to be established, the following elements are required:<br><br>a) The intent of the perpetrator to obtain for themselves or another an unlawful pecuniary benefit;<br><br>b) The knowing misrepresentation of false facts as true, or the unlawful concealment or suppression of true facts, as a result of which—serving as the causal factor—someone is deceived and proceeds to an act, omission, or acquiescence detrimental to th...</code> |
1390
+ | <code>Who can be punished with imprisonment?</code> | <code>1. Anyone who, by knowingly presenting false facts as true or by unlawfully concealing or withholding true facts, damages another person's property by persuading someone to act, omission, or tolerance with the aim of obtaining, for themselves or another, an unlawful financial gain from the damage to that property shall be punished with imprisonment, "and if the damage caused is particularly great, with imprisonment of at least three (3) months and a fine." .<br>If the damage caused exceeds a total of one hundred and twenty thousand (120,000) euros, imprisonment of up to ten (10) years and a fine shall be imposed.<br>2. If the fraud is directed directly against the legal entity of the Greek State, legal entities governed by public law, or local government organizations, and the damage caused exceeds a total of one hundred and twenty thousand (120,000) euros, a prison sentence of at least ten (10) years and a fine of up to one thousand (1,000) daily units shall be imposed. This offense shall b...</code> |
1391
+ * Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters:
1392
+ ```json
1393
+ {
1394
+ "loss": "MultipleNegativesRankingLoss",
1395
+ "matryoshka_dims": [
1396
+ 1024,
1397
+ 768,
1398
+ 512,
1399
+ 256,
1400
+ 128,
1401
+ 64
1402
+ ],
1403
+ "matryoshka_weights": [
1404
+ 1,
1405
+ 1,
1406
+ 1,
1407
+ 1,
1408
+ 1,
1409
+ 1
1410
+ ],
1411
+ "n_dims_per_step": -1
1412
+ }
1413
+ ```
1414
+
1415
+ ### Training Hyperparameters
1416
+ #### Non-Default Hyperparameters
1417
+
1418
+ - `eval_strategy`: epoch
1419
+ - `gradient_accumulation_steps`: 2
1420
+ - `learning_rate`: 2e-05
1421
+ - `num_train_epochs`: 10
1422
+ - `lr_scheduler_type`: cosine
1423
+ - `warmup_ratio`: 0.1
1424
+ - `bf16`: True
1425
+ - `tf32`: True
1426
+ - `load_best_model_at_end`: True
1427
+ - `optim`: adamw_torch_fused
1428
+ - `batch_sampler`: no_duplicates
1429
+
1430
+ #### All Hyperparameters
1431
+ <details><summary>Click to expand</summary>
1432
+
1433
+ - `overwrite_output_dir`: False
1434
+ - `do_predict`: False
1435
+ - `eval_strategy`: epoch
1436
+ - `prediction_loss_only`: True
1437
+ - `per_device_train_batch_size`: 8
1438
+ - `per_device_eval_batch_size`: 8
1439
+ - `per_gpu_train_batch_size`: None
1440
+ - `per_gpu_eval_batch_size`: None
1441
+ - `gradient_accumulation_steps`: 2
1442
+ - `eval_accumulation_steps`: None
1443
+ - `torch_empty_cache_steps`: None
1444
+ - `learning_rate`: 2e-05
1445
+ - `weight_decay`: 0.0
1446
+ - `adam_beta1`: 0.9
1447
+ - `adam_beta2`: 0.999
1448
+ - `adam_epsilon`: 1e-08
1449
+ - `max_grad_norm`: 1.0
1450
+ - `num_train_epochs`: 10
1451
+ - `max_steps`: -1
1452
+ - `lr_scheduler_type`: cosine
1453
+ - `lr_scheduler_kwargs`: {}
1454
+ - `warmup_ratio`: 0.1
1455
+ - `warmup_steps`: 0
1456
+ - `log_level`: passive
1457
+ - `log_level_replica`: warning
1458
+ - `log_on_each_node`: True
1459
+ - `logging_nan_inf_filter`: True
1460
+ - `save_safetensors`: True
1461
+ - `save_on_each_node`: False
1462
+ - `save_only_model`: False
1463
+ - `restore_callback_states_from_checkpoint`: False
1464
+ - `no_cuda`: False
1465
+ - `use_cpu`: False
1466
+ - `use_mps_device`: False
1467
+ - `seed`: 42
1468
+ - `data_seed`: None
1469
+ - `jit_mode_eval`: False
1470
+ - `use_ipex`: False
1471
+ - `bf16`: True
1472
+ - `fp16`: False
1473
+ - `fp16_opt_level`: O1
1474
+ - `half_precision_backend`: auto
1475
+ - `bf16_full_eval`: False
1476
+ - `fp16_full_eval`: False
1477
+ - `tf32`: True
1478
+ - `local_rank`: 0
1479
+ - `ddp_backend`: None
1480
+ - `tpu_num_cores`: None
1481
+ - `tpu_metrics_debug`: False
1482
+ - `debug`: []
1483
+ - `dataloader_drop_last`: False
1484
+ - `dataloader_num_workers`: 0
1485
+ - `dataloader_prefetch_factor`: None
1486
+ - `past_index`: -1
1487
+ - `disable_tqdm`: False
1488
+ - `remove_unused_columns`: True
1489
+ - `label_names`: None
1490
+ - `load_best_model_at_end`: True
1491
+ - `ignore_data_skip`: False
1492
+ - `fsdp`: []
1493
+ - `fsdp_min_num_params`: 0
1494
+ - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
1495
+ - `tp_size`: 0
1496
+ - `fsdp_transformer_layer_cls_to_wrap`: None
1497
+ - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
1498
+ - `deepspeed`: None
1499
+ - `label_smoothing_factor`: 0.0
1500
+ - `optim`: adamw_torch_fused
1501
+ - `optim_args`: None
1502
+ - `adafactor`: False
1503
+ - `group_by_length`: False
1504
+ - `length_column_name`: length
1505
+ - `ddp_find_unused_parameters`: None
1506
+ - `ddp_bucket_cap_mb`: None
1507
+ - `ddp_broadcast_buffers`: False
1508
+ - `dataloader_pin_memory`: True
1509
+ - `dataloader_persistent_workers`: False
1510
+ - `skip_memory_metrics`: True
1511
+ - `use_legacy_prediction_loop`: False
1512
+ - `push_to_hub`: False
1513
+ - `resume_from_checkpoint`: None
1514
+ - `hub_model_id`: None
1515
+ - `hub_strategy`: every_save
1516
+ - `hub_private_repo`: None
1517
+ - `hub_always_push`: False
1518
+ - `gradient_checkpointing`: False
1519
+ - `gradient_checkpointing_kwargs`: None
1520
+ - `include_inputs_for_metrics`: False
1521
+ - `include_for_metrics`: []
1522
+ - `eval_do_concat_batches`: True
1523
+ - `fp16_backend`: auto
1524
+ - `push_to_hub_model_id`: None
1525
+ - `push_to_hub_organization`: None
1526
+ - `mp_parameters`:
1527
+ - `auto_find_batch_size`: False
1528
+ - `full_determinism`: False
1529
+ - `torchdynamo`: None
1530
+ - `ray_scope`: last
1531
+ - `ddp_timeout`: 1800
1532
+ - `torch_compile`: False
1533
+ - `torch_compile_backend`: None
1534
+ - `torch_compile_mode`: None
1535
+ - `include_tokens_per_second`: False
1536
+ - `include_num_input_tokens_seen`: False
1537
+ - `neftune_noise_alpha`: None
1538
+ - `optim_target_modules`: None
1539
+ - `batch_eval_metrics`: False
1540
+ - `eval_on_start`: False
1541
+ - `use_liger_kernel`: False
1542
+ - `eval_use_gather_object`: False
1543
+ - `average_tokens_across_devices`: False
1544
+ - `prompts`: None
1545
+ - `batch_sampler`: no_duplicates
1546
+ - `multi_dataset_batch_sampler`: proportional
1547
+ - `router_mapping`: {}
1548
+ - `learning_rate_mapping`: {}
1549
+
1550
+ </details>
1551
+
1552
+ ### Training Logs
1553
+ | Epoch | Step | Training Loss | dim_1024_cosine_ndcg@10 | dim_768_cosine_ndcg@10 | dim_512_cosine_ndcg@10 | dim_256_cosine_ndcg@10 | dim_128_cosine_ndcg@10 | dim_64_cosine_ndcg@10 |
1554
+ |:------:|:----:|:-------------:|:-----------------------:|:----------------------:|:----------------------:|:----------------------:|:----------------------:|:---------------------:|
1555
+ | 0.1818 | 1 | 18.029 | - | - | - | - | - | - |
1556
+ | 0.3636 | 2 | 19.4106 | - | - | - | - | - | - |
1557
+ | 0.5455 | 3 | 16.6201 | - | - | - | - | - | - |
1558
+ | 0.7273 | 4 | 15.3048 | - | - | - | - | - | - |
1559
+ | 0.9091 | 5 | 14.0182 | - | - | - | - | - | - |
1560
+ | 1.0 | 6 | 6.4771 | - | - | - | - | - | - |
1561
+ | 1.0909 | 7 | 6.7664 | 0.6167 | 0.5821 | 0.5524 | 0.5177 | 0.5278 | 0.4124 |
1562
+ | 1.1818 | 8 | 11.8583 | - | - | - | - | - | - |
1563
+ | 1.3636 | 9 | 11.9216 | - | - | - | - | - | - |
1564
+ | 1.5455 | 10 | 13.3764 | - | - | - | - | - | - |
1565
+ | 1.7273 | 11 | 12.9063 | - | - | - | - | - | - |
1566
+ | 1.9091 | 12 | 13.5984 | - | - | - | - | - | - |
1567
+ | 2.0 | 13 | 7.8523 | - | - | - | - | - | - |
1568
+ | 2.0909 | 14 | 4.4487 | 0.5921 | 0.5921 | 0.5518 | 0.5709 | 0.5685 | 0.5113 |
1569
+ | 2.1818 | 15 | 8.5374 | - | - | - | - | - | - |
1570
+ | 2.3636 | 16 | 9.6999 | - | - | - | - | - | - |
1571
+ | 2.5455 | 17 | 9.0121 | - | - | - | - | - | - |
1572
+ | 2.7273 | 18 | 13.5705 | - | - | - | - | - | - |
1573
+ | 2.9091 | 19 | 13.0195 | - | - | - | - | - | - |
1574
+ | 3.0 | 20 | 7.9821 | - | - | - | - | - | - |
1575
+ | 3.0909 | 21 | 3.2842 | 0.5159 | 0.5636 | 0.5468 | 0.5468 | 0.5468 | 0.5233 |
1576
+ | 3.1818 | 22 | 4.4446 | - | - | - | - | - | - |
1577
+ | 3.3636 | 23 | 5.7244 | - | - | - | - | - | - |
1578
+ | 3.5455 | 24 | 7.1394 | - | - | - | - | - | - |
1579
+ | 3.7273 | 25 | 16.7583 | - | - | - | - | - | - |
1580
+ | 3.9091 | 26 | 11.3515 | - | - | - | - | - | - |
1581
+ | 4.0 | 27 | 8.813 | - | - | - | - | - | - |
1582
+ | 4.0909 | 28 | 6.9124 | 0.5159 | 0.5468 | 0.4992 | 0.5468 | 0.4992 | 0.4992 |
1583
+ | 4.1818 | 29 | 6.1814 | - | - | - | - | - | - |
1584
+ | 4.3636 | 30 | 7.1606 | - | - | - | - | - | - |
1585
+ | 4.5455 | 31 | 5.0888 | - | - | - | - | - | - |
1586
+ | 4.7273 | 32 | 5.0684 | - | - | - | - | - | - |
1587
+ | 4.9091 | 33 | 6.7382 | - | - | - | - | - | - |
1588
+ | 5.0 | 34 | 7.0497 | - | - | - | - | - | - |
1589
+ | 5.0909 | 35 | 6.582 | 0.5598 | 0.5598 | 0.5598 | 0.6421 | 0.5598 | 0.4918 |
1590
+
1591
+
1592
+ ### Framework Versions
1593
+ - Python: 3.12.12
1594
+ - Sentence Transformers: 5.1.1
1595
+ - Transformers: 4.51.3
1596
+ - PyTorch: 2.8.0+cu126
1597
+ - Accelerate: 1.11.0
1598
+ - Datasets: 4.0.0
1599
+ - Tokenizers: 0.21.4
1600
+
1601
+ ## Citation
1602
+
1603
+ ### BibTeX
1604
+
1605
+ #### Sentence Transformers
1606
+ ```bibtex
1607
+ @inproceedings{reimers-2019-sentence-bert,
1608
+ title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
1609
+ author = "Reimers, Nils and Gurevych, Iryna",
1610
+ booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
1611
+ month = "11",
1612
+ year = "2019",
1613
+ publisher = "Association for Computational Linguistics",
1614
+ url = "https://arxiv.org/abs/1908.10084",
1615
+ }
1616
+ ```
1617
+
1618
+ #### MatryoshkaLoss
1619
+ ```bibtex
1620
+ @misc{kusupati2024matryoshka,
1621
+ title={Matryoshka Representation Learning},
1622
+ 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},
1623
+ year={2024},
1624
+ eprint={2205.13147},
1625
+ archivePrefix={arXiv},
1626
+ primaryClass={cs.LG}
1627
+ }
1628
+ ```
1629
+
1630
+ #### MultipleNegativesRankingLoss
1631
+ ```bibtex
1632
+ @misc{henderson2017efficient,
1633
+ title={Efficient Natural Language Response Suggestion for Smart Reply},
1634
+ 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},
1635
+ year={2017},
1636
+ eprint={1705.00652},
1637
+ archivePrefix={arXiv},
1638
+ primaryClass={cs.CL}
1639
+ }
1640
+ ```
1641
+
1642
+ <!--
1643
+ ## Glossary
1644
+
1645
+ *Clearly define terms in order to be accessible across audiences.*
1646
+ -->
1647
+
1648
+ <!--
1649
+ ## Model Card Authors
1650
+
1651
+ *Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
1652
+ -->
1653
+
1654
+ <!--
1655
+ ## Model Card Contact
1656
+
1657
+ *Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
1658
+ -->
checkpoint-35/config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "XLMRobertaModel"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 0,
7
+ "classifier_dropout": null,
8
+ "eos_token_id": 2,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 1024,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 4096,
14
+ "layer_norm_eps": 1e-05,
15
+ "max_position_embeddings": 514,
16
+ "model_type": "xlm-roberta",
17
+ "num_attention_heads": 16,
18
+ "num_hidden_layers": 24,
19
+ "output_past": true,
20
+ "pad_token_id": 1,
21
+ "position_embedding_type": "absolute",
22
+ "torch_dtype": "float32",
23
+ "transformers_version": "4.51.3",
24
+ "type_vocab_size": 1,
25
+ "use_cache": true,
26
+ "vocab_size": 250002
27
+ }
checkpoint-35/config_sentence_transformers.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "SentenceTransformer",
3
+ "__version__": {
4
+ "sentence_transformers": "5.1.1",
5
+ "transformers": "4.51.3",
6
+ "pytorch": "2.8.0+cu126"
7
+ },
8
+ "prompts": {
9
+ "query": "",
10
+ "document": ""
11
+ },
12
+ "default_prompt_name": null,
13
+ "similarity_fn_name": "cosine"
14
+ }
checkpoint-35/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f7897be8f9d594942e4a3af80d3943c8b741feec27f8307cc90b80b9524898c
3
+ size 2239607176
checkpoint-35/modules.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "idx": 0,
4
+ "name": "0",
5
+ "path": "",
6
+ "type": "sentence_transformers.models.Transformer"
7
+ },
8
+ {
9
+ "idx": 1,
10
+ "name": "1",
11
+ "path": "1_Pooling",
12
+ "type": "sentence_transformers.models.Pooling"
13
+ },
14
+ {
15
+ "idx": 2,
16
+ "name": "2",
17
+ "path": "2_Normalize",
18
+ "type": "sentence_transformers.models.Normalize"
19
+ }
20
+ ]
checkpoint-35/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c026b136a1e7d5fc0d68e1f5c9885274ee96bd6b327534f96178dde6c62f3d9
3
+ size 4471067142
checkpoint-35/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c2e5f9980b1d32c44b03ed896bf9b5d0a406317fc3796843a35decde00b68a3b
3
+ size 14645
checkpoint-35/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a649381ddd14598478361c263c6e25680ae024d7e4ac70d033eba4fa322d038
3
+ size 1465
checkpoint-35/sentence_bert_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "max_seq_length": 512,
3
+ "do_lower_case": false
4
+ }
checkpoint-35/sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
3
+ size 5069051
checkpoint-35/special_tokens_map.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "cls_token": {
10
+ "content": "<s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "eos_token": {
17
+ "content": "</s>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "mask_token": {
24
+ "content": "<mask>",
25
+ "lstrip": true,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "pad_token": {
31
+ "content": "<pad>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ },
37
+ "sep_token": {
38
+ "content": "</s>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false
43
+ },
44
+ "unk_token": {
45
+ "content": "<unk>",
46
+ "lstrip": false,
47
+ "normalized": false,
48
+ "rstrip": false,
49
+ "single_word": false
50
+ }
51
+ }
checkpoint-35/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:883b037111086fd4dfebbbc9b7cee11e1517b5e0c0514879478661440f137085
3
+ size 17082987
checkpoint-35/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "250001": {
36
+ "content": "<mask>",
37
+ "lstrip": true,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "bos_token": "<s>",
45
+ "clean_up_tokenization_spaces": true,
46
+ "cls_token": "<s>",
47
+ "eos_token": "</s>",
48
+ "extra_special_tokens": {},
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 512,
51
+ "pad_token": "<pad>",
52
+ "sep_token": "</s>",
53
+ "tokenizer_class": "XLMRobertaTokenizer",
54
+ "unk_token": "<unk>"
55
+ }
checkpoint-35/trainer_state.json ADDED
@@ -0,0 +1,778 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 14,
3
+ "best_metric": 0.5685354415901852,
4
+ "best_model_checkpoint": "multilingual-e5-large/checkpoint-14",
5
+ "epoch": 5.090909090909091,
6
+ "eval_steps": 500,
7
+ "global_step": 35,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.18181818181818182,
14
+ "grad_norm": 135.6220703125,
15
+ "learning_rate": 0.0,
16
+ "loss": 18.029,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.36363636363636365,
21
+ "grad_norm": 140.5317840576172,
22
+ "learning_rate": 4.000000000000001e-06,
23
+ "loss": 19.4106,
24
+ "step": 2
25
+ },
26
+ {
27
+ "epoch": 0.5454545454545454,
28
+ "grad_norm": 142.2399444580078,
29
+ "learning_rate": 8.000000000000001e-06,
30
+ "loss": 16.6201,
31
+ "step": 3
32
+ },
33
+ {
34
+ "epoch": 0.7272727272727273,
35
+ "grad_norm": 110.63945770263672,
36
+ "learning_rate": 1.2e-05,
37
+ "loss": 15.3048,
38
+ "step": 4
39
+ },
40
+ {
41
+ "epoch": 0.9090909090909091,
42
+ "grad_norm": 129.43089294433594,
43
+ "learning_rate": 1.6000000000000003e-05,
44
+ "loss": 14.0182,
45
+ "step": 5
46
+ },
47
+ {
48
+ "epoch": 1.0,
49
+ "grad_norm": 108.25520324707031,
50
+ "learning_rate": 2e-05,
51
+ "loss": 6.4771,
52
+ "step": 6
53
+ },
54
+ {
55
+ "epoch": 1.0909090909090908,
56
+ "grad_norm": 95.05782318115234,
57
+ "learning_rate": 1.9975640502598243e-05,
58
+ "loss": 6.7664,
59
+ "step": 7
60
+ },
61
+ {
62
+ "epoch": 1.0909090909090908,
63
+ "eval_dim_1024_cosine_accuracy@1": 0.5714285714285714,
64
+ "eval_dim_1024_cosine_accuracy@10": 0.6666666666666666,
65
+ "eval_dim_1024_cosine_accuracy@3": 0.5714285714285714,
66
+ "eval_dim_1024_cosine_accuracy@5": 0.6190476190476191,
67
+ "eval_dim_1024_cosine_map@100": 0.6946498352795571,
68
+ "eval_dim_1024_cosine_mrr@10": 0.5888888888888888,
69
+ "eval_dim_1024_cosine_ndcg@10": 0.6167053425672016,
70
+ "eval_dim_1024_cosine_precision@1": 0.5714285714285714,
71
+ "eval_dim_1024_cosine_precision@10": 0.4476190476190477,
72
+ "eval_dim_1024_cosine_precision@3": 0.5555555555555556,
73
+ "eval_dim_1024_cosine_precision@5": 0.5142857142857142,
74
+ "eval_dim_1024_cosine_recall@1": 0.10123626373626372,
75
+ "eval_dim_1024_cosine_recall@10": 0.5401404151404151,
76
+ "eval_dim_1024_cosine_recall@3": 0.2798992673992674,
77
+ "eval_dim_1024_cosine_recall@5": 0.3871336996336997,
78
+ "eval_dim_128_cosine_accuracy@1": 0.47619047619047616,
79
+ "eval_dim_128_cosine_accuracy@10": 0.5714285714285714,
80
+ "eval_dim_128_cosine_accuracy@3": 0.47619047619047616,
81
+ "eval_dim_128_cosine_accuracy@5": 0.5238095238095238,
82
+ "eval_dim_128_cosine_map@100": 0.5954946117864774,
83
+ "eval_dim_128_cosine_mrr@10": 0.49251700680272104,
84
+ "eval_dim_128_cosine_ndcg@10": 0.5277801377642622,
85
+ "eval_dim_128_cosine_precision@1": 0.47619047619047616,
86
+ "eval_dim_128_cosine_precision@10": 0.3761904761904762,
87
+ "eval_dim_128_cosine_precision@3": 0.4603174603174603,
88
+ "eval_dim_128_cosine_precision@5": 0.42857142857142855,
89
+ "eval_dim_128_cosine_recall@1": 0.08536324786324785,
90
+ "eval_dim_128_cosine_recall@10": 0.4806166056166057,
91
+ "eval_dim_128_cosine_recall@3": 0.23228021978021982,
92
+ "eval_dim_128_cosine_recall@5": 0.31967338217338215,
93
+ "eval_dim_256_cosine_accuracy@1": 0.47619047619047616,
94
+ "eval_dim_256_cosine_accuracy@10": 0.5714285714285714,
95
+ "eval_dim_256_cosine_accuracy@3": 0.47619047619047616,
96
+ "eval_dim_256_cosine_accuracy@5": 0.5238095238095238,
97
+ "eval_dim_256_cosine_map@100": 0.6218563539505393,
98
+ "eval_dim_256_cosine_mrr@10": 0.4925170068027211,
99
+ "eval_dim_256_cosine_ndcg@10": 0.5177339964253599,
100
+ "eval_dim_256_cosine_precision@1": 0.47619047619047616,
101
+ "eval_dim_256_cosine_precision@10": 0.3476190476190476,
102
+ "eval_dim_256_cosine_precision@3": 0.4603174603174603,
103
+ "eval_dim_256_cosine_precision@5": 0.419047619047619,
104
+ "eval_dim_256_cosine_recall@1": 0.09543650793650793,
105
+ "eval_dim_256_cosine_recall@10": 0.47817460317460325,
106
+ "eval_dim_256_cosine_recall@3": 0.2625,
107
+ "eval_dim_256_cosine_recall@5": 0.35813492063492064,
108
+ "eval_dim_512_cosine_accuracy@1": 0.5238095238095238,
109
+ "eval_dim_512_cosine_accuracy@10": 0.5714285714285714,
110
+ "eval_dim_512_cosine_accuracy@3": 0.5238095238095238,
111
+ "eval_dim_512_cosine_accuracy@5": 0.5714285714285714,
112
+ "eval_dim_512_cosine_map@100": 0.6608463504287858,
113
+ "eval_dim_512_cosine_mrr@10": 0.5333333333333333,
114
+ "eval_dim_512_cosine_ndcg@10": 0.5523687509163372,
115
+ "eval_dim_512_cosine_precision@1": 0.5238095238095238,
116
+ "eval_dim_512_cosine_precision@10": 0.3761904761904762,
117
+ "eval_dim_512_cosine_precision@3": 0.5079365079365079,
118
+ "eval_dim_512_cosine_precision@5": 0.4666666666666666,
119
+ "eval_dim_512_cosine_recall@1": 0.0994047619047619,
120
+ "eval_dim_512_cosine_recall@10": 0.5019841269841271,
121
+ "eval_dim_512_cosine_recall@3": 0.2744047619047619,
122
+ "eval_dim_512_cosine_recall@5": 0.37797619047619047,
123
+ "eval_dim_64_cosine_accuracy@1": 0.3333333333333333,
124
+ "eval_dim_64_cosine_accuracy@10": 0.5714285714285714,
125
+ "eval_dim_64_cosine_accuracy@3": 0.38095238095238093,
126
+ "eval_dim_64_cosine_accuracy@5": 0.38095238095238093,
127
+ "eval_dim_64_cosine_map@100": 0.5579595177809107,
128
+ "eval_dim_64_cosine_mrr@10": 0.3764172335600907,
129
+ "eval_dim_64_cosine_ndcg@10": 0.41244392103396355,
130
+ "eval_dim_64_cosine_precision@1": 0.3333333333333333,
131
+ "eval_dim_64_cosine_precision@10": 0.2619047619047619,
132
+ "eval_dim_64_cosine_precision@3": 0.3492063492063492,
133
+ "eval_dim_64_cosine_precision@5": 0.3142857142857143,
134
+ "eval_dim_64_cosine_recall@1": 0.07063492063492063,
135
+ "eval_dim_64_cosine_recall@10": 0.42002442002442003,
136
+ "eval_dim_64_cosine_recall@3": 0.2357142857142857,
137
+ "eval_dim_64_cosine_recall@5": 0.34523809523809523,
138
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
139
+ "eval_dim_768_cosine_accuracy@10": 0.6666666666666666,
140
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
141
+ "eval_dim_768_cosine_accuracy@5": 0.5714285714285714,
142
+ "eval_dim_768_cosine_map@100": 0.6727109022414055,
143
+ "eval_dim_768_cosine_mrr@10": 0.5480725623582765,
144
+ "eval_dim_768_cosine_ndcg@10": 0.5820705880762242,
145
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
146
+ "eval_dim_768_cosine_precision@10": 0.41904761904761906,
147
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
148
+ "eval_dim_768_cosine_precision@5": 0.4666666666666666,
149
+ "eval_dim_768_cosine_recall@1": 0.09726800976800977,
150
+ "eval_dim_768_cosine_recall@10": 0.5163308913308913,
151
+ "eval_dim_768_cosine_recall@3": 0.2679945054945055,
152
+ "eval_dim_768_cosine_recall@5": 0.36729242979242976,
153
+ "eval_runtime": 7.0357,
154
+ "eval_samples_per_second": 0.0,
155
+ "eval_sequential_score": 0.41244392103396355,
156
+ "eval_steps_per_second": 0.0,
157
+ "step": 7
158
+ },
159
+ {
160
+ "epoch": 1.1818181818181819,
161
+ "grad_norm": 159.73828125,
162
+ "learning_rate": 1.9902680687415704e-05,
163
+ "loss": 11.8583,
164
+ "step": 8
165
+ },
166
+ {
167
+ "epoch": 1.3636363636363638,
168
+ "grad_norm": 134.70603942871094,
169
+ "learning_rate": 1.9781476007338058e-05,
170
+ "loss": 11.9216,
171
+ "step": 9
172
+ },
173
+ {
174
+ "epoch": 1.5454545454545454,
175
+ "grad_norm": 139.73654174804688,
176
+ "learning_rate": 1.961261695938319e-05,
177
+ "loss": 13.3764,
178
+ "step": 10
179
+ },
180
+ {
181
+ "epoch": 1.7272727272727273,
182
+ "grad_norm": 173.23065185546875,
183
+ "learning_rate": 1.9396926207859085e-05,
184
+ "loss": 12.9063,
185
+ "step": 11
186
+ },
187
+ {
188
+ "epoch": 1.9090909090909092,
189
+ "grad_norm": 208.78482055664062,
190
+ "learning_rate": 1.913545457642601e-05,
191
+ "loss": 13.5984,
192
+ "step": 12
193
+ },
194
+ {
195
+ "epoch": 2.0,
196
+ "grad_norm": 298.82183837890625,
197
+ "learning_rate": 1.8829475928589272e-05,
198
+ "loss": 7.8523,
199
+ "step": 13
200
+ },
201
+ {
202
+ "epoch": 2.090909090909091,
203
+ "grad_norm": 156.5718994140625,
204
+ "learning_rate": 1.848048096156426e-05,
205
+ "loss": 4.4487,
206
+ "step": 14
207
+ },
208
+ {
209
+ "epoch": 2.090909090909091,
210
+ "eval_dim_1024_cosine_accuracy@1": 0.5238095238095238,
211
+ "eval_dim_1024_cosine_accuracy@10": 0.6666666666666666,
212
+ "eval_dim_1024_cosine_accuracy@3": 0.5238095238095238,
213
+ "eval_dim_1024_cosine_accuracy@5": 0.5714285714285714,
214
+ "eval_dim_1024_cosine_map@100": 0.67423207909377,
215
+ "eval_dim_1024_cosine_mrr@10": 0.5480725623582765,
216
+ "eval_dim_1024_cosine_ndcg@10": 0.5921167294151266,
217
+ "eval_dim_1024_cosine_precision@1": 0.5238095238095238,
218
+ "eval_dim_1024_cosine_precision@10": 0.4476190476190477,
219
+ "eval_dim_1024_cosine_precision@3": 0.5079365079365079,
220
+ "eval_dim_1024_cosine_precision@5": 0.47619047619047616,
221
+ "eval_dim_1024_cosine_recall@1": 0.08933150183150182,
222
+ "eval_dim_1024_cosine_recall@10": 0.5401404151404151,
223
+ "eval_dim_1024_cosine_recall@3": 0.24418498168498168,
224
+ "eval_dim_1024_cosine_recall@5": 0.33951465201465203,
225
+ "eval_dim_128_cosine_accuracy@1": 0.5238095238095238,
226
+ "eval_dim_128_cosine_accuracy@10": 0.6190476190476191,
227
+ "eval_dim_128_cosine_accuracy@3": 0.5238095238095238,
228
+ "eval_dim_128_cosine_accuracy@5": 0.5714285714285714,
229
+ "eval_dim_128_cosine_map@100": 0.6489604480560528,
230
+ "eval_dim_128_cosine_mrr@10": 0.5401360544217686,
231
+ "eval_dim_128_cosine_ndcg@10": 0.5685354415901852,
232
+ "eval_dim_128_cosine_precision@1": 0.5238095238095238,
233
+ "eval_dim_128_cosine_precision@10": 0.4238095238095239,
234
+ "eval_dim_128_cosine_precision@3": 0.5238095238095238,
235
+ "eval_dim_128_cosine_precision@5": 0.5047619047619047,
236
+ "eval_dim_128_cosine_recall@1": 0.07345848595848595,
237
+ "eval_dim_128_cosine_recall@10": 0.5202991452991453,
238
+ "eval_dim_128_cosine_recall@3": 0.2203754578754579,
239
+ "eval_dim_128_cosine_recall@5": 0.34745115995116,
240
+ "eval_dim_256_cosine_accuracy@1": 0.5238095238095238,
241
+ "eval_dim_256_cosine_accuracy@10": 0.6190476190476191,
242
+ "eval_dim_256_cosine_accuracy@3": 0.5238095238095238,
243
+ "eval_dim_256_cosine_accuracy@5": 0.5714285714285714,
244
+ "eval_dim_256_cosine_map@100": 0.651530364911684,
245
+ "eval_dim_256_cosine_mrr@10": 0.5401360544217686,
246
+ "eval_dim_256_cosine_ndcg@10": 0.5708936958722651,
247
+ "eval_dim_256_cosine_precision@1": 0.5238095238095238,
248
+ "eval_dim_256_cosine_precision@10": 0.4238095238095239,
249
+ "eval_dim_256_cosine_precision@3": 0.5079365079365079,
250
+ "eval_dim_256_cosine_precision@5": 0.49523809523809514,
251
+ "eval_dim_256_cosine_recall@1": 0.0813949938949939,
252
+ "eval_dim_256_cosine_recall@10": 0.5202991452991453,
253
+ "eval_dim_256_cosine_recall@3": 0.22037545787545787,
254
+ "eval_dim_256_cosine_recall@5": 0.33951465201465203,
255
+ "eval_dim_512_cosine_accuracy@1": 0.47619047619047616,
256
+ "eval_dim_512_cosine_accuracy@10": 0.6190476190476191,
257
+ "eval_dim_512_cosine_accuracy@3": 0.47619047619047616,
258
+ "eval_dim_512_cosine_accuracy@5": 0.5714285714285714,
259
+ "eval_dim_512_cosine_map@100": 0.6265911712939339,
260
+ "eval_dim_512_cosine_mrr@10": 0.5020408163265305,
261
+ "eval_dim_512_cosine_ndcg@10": 0.5518338753600308,
262
+ "eval_dim_512_cosine_precision@1": 0.47619047619047616,
263
+ "eval_dim_512_cosine_precision@10": 0.4238095238095239,
264
+ "eval_dim_512_cosine_precision@3": 0.4603174603174603,
265
+ "eval_dim_512_cosine_precision@5": 0.45714285714285713,
266
+ "eval_dim_512_cosine_recall@1": 0.07345848595848595,
267
+ "eval_dim_512_cosine_recall@10": 0.5202991452991453,
268
+ "eval_dim_512_cosine_recall@3": 0.19656593406593406,
269
+ "eval_dim_512_cosine_recall@5": 0.3077686202686203,
270
+ "eval_dim_64_cosine_accuracy@1": 0.42857142857142855,
271
+ "eval_dim_64_cosine_accuracy@10": 0.6190476190476191,
272
+ "eval_dim_64_cosine_accuracy@3": 0.42857142857142855,
273
+ "eval_dim_64_cosine_accuracy@5": 0.47619047619047616,
274
+ "eval_dim_64_cosine_map@100": 0.5888462989137369,
275
+ "eval_dim_64_cosine_mrr@10": 0.45963718820861665,
276
+ "eval_dim_64_cosine_ndcg@10": 0.51131642091388,
277
+ "eval_dim_64_cosine_precision@1": 0.42857142857142855,
278
+ "eval_dim_64_cosine_precision@10": 0.3999999999999999,
279
+ "eval_dim_64_cosine_precision@3": 0.42857142857142855,
280
+ "eval_dim_64_cosine_precision@5": 0.42857142857142855,
281
+ "eval_dim_64_cosine_recall@1": 0.053617216117216114,
282
+ "eval_dim_64_cosine_recall@10": 0.5004578754578755,
283
+ "eval_dim_64_cosine_recall@3": 0.16085164835164836,
284
+ "eval_dim_64_cosine_recall@5": 0.27205433455433453,
285
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
286
+ "eval_dim_768_cosine_accuracy@10": 0.6666666666666666,
287
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
288
+ "eval_dim_768_cosine_accuracy@5": 0.5714285714285714,
289
+ "eval_dim_768_cosine_map@100": 0.67423207909377,
290
+ "eval_dim_768_cosine_mrr@10": 0.5480725623582765,
291
+ "eval_dim_768_cosine_ndcg@10": 0.5921167294151266,
292
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
293
+ "eval_dim_768_cosine_precision@10": 0.4476190476190477,
294
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
295
+ "eval_dim_768_cosine_precision@5": 0.47619047619047616,
296
+ "eval_dim_768_cosine_recall@1": 0.08933150183150182,
297
+ "eval_dim_768_cosine_recall@10": 0.5401404151404151,
298
+ "eval_dim_768_cosine_recall@3": 0.24418498168498168,
299
+ "eval_dim_768_cosine_recall@5": 0.33951465201465203,
300
+ "eval_runtime": 6.9723,
301
+ "eval_samples_per_second": 0.0,
302
+ "eval_sequential_score": 0.51131642091388,
303
+ "eval_steps_per_second": 0.0,
304
+ "step": 14
305
+ },
306
+ {
307
+ "epoch": 2.1818181818181817,
308
+ "grad_norm": 156.31887817382812,
309
+ "learning_rate": 1.8090169943749477e-05,
310
+ "loss": 8.5374,
311
+ "step": 15
312
+ },
313
+ {
314
+ "epoch": 2.3636363636363638,
315
+ "grad_norm": 205.52923583984375,
316
+ "learning_rate": 1.766044443118978e-05,
317
+ "loss": 9.6999,
318
+ "step": 16
319
+ },
320
+ {
321
+ "epoch": 2.5454545454545454,
322
+ "grad_norm": 254.7772216796875,
323
+ "learning_rate": 1.7193398003386514e-05,
324
+ "loss": 9.0121,
325
+ "step": 17
326
+ },
327
+ {
328
+ "epoch": 2.7272727272727275,
329
+ "grad_norm": 476.75164794921875,
330
+ "learning_rate": 1.6691306063588583e-05,
331
+ "loss": 13.5705,
332
+ "step": 18
333
+ },
334
+ {
335
+ "epoch": 2.909090909090909,
336
+ "grad_norm": 376.09954833984375,
337
+ "learning_rate": 1.6156614753256583e-05,
338
+ "loss": 13.0195,
339
+ "step": 19
340
+ },
341
+ {
342
+ "epoch": 3.0,
343
+ "grad_norm": 387.24322509765625,
344
+ "learning_rate": 1.5591929034707468e-05,
345
+ "loss": 7.9821,
346
+ "step": 20
347
+ },
348
+ {
349
+ "epoch": 3.090909090909091,
350
+ "grad_norm": 210.42813110351562,
351
+ "learning_rate": 1.5000000000000002e-05,
352
+ "loss": 3.2842,
353
+ "step": 21
354
+ },
355
+ {
356
+ "epoch": 3.090909090909091,
357
+ "eval_dim_1024_cosine_accuracy@1": 0.47619047619047616,
358
+ "eval_dim_1024_cosine_accuracy@10": 0.5714285714285714,
359
+ "eval_dim_1024_cosine_accuracy@3": 0.47619047619047616,
360
+ "eval_dim_1024_cosine_accuracy@5": 0.47619047619047616,
361
+ "eval_dim_1024_cosine_map@100": 0.616370861587526,
362
+ "eval_dim_1024_cosine_mrr@10": 0.49092970521541945,
363
+ "eval_dim_1024_cosine_ndcg@10": 0.5159384546892658,
364
+ "eval_dim_1024_cosine_precision@1": 0.47619047619047616,
365
+ "eval_dim_1024_cosine_precision@10": 0.4,
366
+ "eval_dim_1024_cosine_precision@3": 0.4603174603174603,
367
+ "eval_dim_1024_cosine_precision@5": 0.419047619047619,
368
+ "eval_dim_1024_cosine_recall@1": 0.07822039072039072,
369
+ "eval_dim_1024_cosine_recall@10": 0.4449023199023199,
370
+ "eval_dim_1024_cosine_recall@3": 0.21085164835164832,
371
+ "eval_dim_1024_cosine_recall@5": 0.27602258852258854,
372
+ "eval_dim_128_cosine_accuracy@1": 0.5238095238095238,
373
+ "eval_dim_128_cosine_accuracy@10": 0.5714285714285714,
374
+ "eval_dim_128_cosine_accuracy@3": 0.5238095238095238,
375
+ "eval_dim_128_cosine_accuracy@5": 0.5238095238095238,
376
+ "eval_dim_128_cosine_map@100": 0.6278310491545506,
377
+ "eval_dim_128_cosine_mrr@10": 0.5306122448979591,
378
+ "eval_dim_128_cosine_ndcg@10": 0.5468399582764966,
379
+ "eval_dim_128_cosine_precision@1": 0.5238095238095238,
380
+ "eval_dim_128_cosine_precision@10": 0.4238095238095239,
381
+ "eval_dim_128_cosine_precision@3": 0.5079365079365079,
382
+ "eval_dim_128_cosine_precision@5": 0.4666666666666666,
383
+ "eval_dim_128_cosine_recall@1": 0.0811965811965812,
384
+ "eval_dim_128_cosine_recall@10": 0.4548229548229548,
385
+ "eval_dim_128_cosine_recall@3": 0.21978021978021975,
386
+ "eval_dim_128_cosine_recall@5": 0.2909035409035409,
387
+ "eval_dim_256_cosine_accuracy@1": 0.5238095238095238,
388
+ "eval_dim_256_cosine_accuracy@10": 0.5714285714285714,
389
+ "eval_dim_256_cosine_accuracy@3": 0.5238095238095238,
390
+ "eval_dim_256_cosine_accuracy@5": 0.5238095238095238,
391
+ "eval_dim_256_cosine_map@100": 0.6295752260160459,
392
+ "eval_dim_256_cosine_mrr@10": 0.5306122448979591,
393
+ "eval_dim_256_cosine_ndcg@10": 0.5468399582764966,
394
+ "eval_dim_256_cosine_precision@1": 0.5238095238095238,
395
+ "eval_dim_256_cosine_precision@10": 0.4238095238095239,
396
+ "eval_dim_256_cosine_precision@3": 0.5079365079365079,
397
+ "eval_dim_256_cosine_precision@5": 0.4666666666666666,
398
+ "eval_dim_256_cosine_recall@1": 0.0811965811965812,
399
+ "eval_dim_256_cosine_recall@10": 0.4548229548229548,
400
+ "eval_dim_256_cosine_recall@3": 0.21978021978021975,
401
+ "eval_dim_256_cosine_recall@5": 0.2909035409035409,
402
+ "eval_dim_512_cosine_accuracy@1": 0.5238095238095238,
403
+ "eval_dim_512_cosine_accuracy@10": 0.5714285714285714,
404
+ "eval_dim_512_cosine_accuracy@3": 0.5238095238095238,
405
+ "eval_dim_512_cosine_accuracy@5": 0.5238095238095238,
406
+ "eval_dim_512_cosine_map@100": 0.6298362122328386,
407
+ "eval_dim_512_cosine_mrr@10": 0.5306122448979591,
408
+ "eval_dim_512_cosine_ndcg@10": 0.5468399582764966,
409
+ "eval_dim_512_cosine_precision@1": 0.5238095238095238,
410
+ "eval_dim_512_cosine_precision@10": 0.4238095238095239,
411
+ "eval_dim_512_cosine_precision@3": 0.5079365079365079,
412
+ "eval_dim_512_cosine_precision@5": 0.4666666666666666,
413
+ "eval_dim_512_cosine_recall@1": 0.0811965811965812,
414
+ "eval_dim_512_cosine_recall@10": 0.4548229548229548,
415
+ "eval_dim_512_cosine_recall@3": 0.21978021978021975,
416
+ "eval_dim_512_cosine_recall@5": 0.2909035409035409,
417
+ "eval_dim_64_cosine_accuracy@1": 0.47619047619047616,
418
+ "eval_dim_64_cosine_accuracy@10": 0.5714285714285714,
419
+ "eval_dim_64_cosine_accuracy@3": 0.47619047619047616,
420
+ "eval_dim_64_cosine_accuracy@5": 0.5238095238095238,
421
+ "eval_dim_64_cosine_map@100": 0.6058793096030948,
422
+ "eval_dim_64_cosine_mrr@10": 0.49251700680272104,
423
+ "eval_dim_64_cosine_ndcg@10": 0.5232746482532176,
424
+ "eval_dim_64_cosine_precision@1": 0.47619047619047616,
425
+ "eval_dim_64_cosine_precision@10": 0.4238095238095238,
426
+ "eval_dim_64_cosine_precision@3": 0.4603174603174603,
427
+ "eval_dim_64_cosine_precision@5": 0.4476190476190475,
428
+ "eval_dim_64_cosine_recall@1": 0.06631562881562882,
429
+ "eval_dim_64_cosine_recall@10": 0.46474358974358976,
430
+ "eval_dim_64_cosine_recall@3": 0.17513736263736263,
431
+ "eval_dim_64_cosine_recall@5": 0.2641178266178266,
432
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
433
+ "eval_dim_768_cosine_accuracy@10": 0.6190476190476191,
434
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
435
+ "eval_dim_768_cosine_accuracy@5": 0.5238095238095238,
436
+ "eval_dim_768_cosine_map@100": 0.6486454426450273,
437
+ "eval_dim_768_cosine_mrr@10": 0.538548752834467,
438
+ "eval_dim_768_cosine_ndcg@10": 0.5635575023083134,
439
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
440
+ "eval_dim_768_cosine_precision@10": 0.4476190476190477,
441
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
442
+ "eval_dim_768_cosine_precision@5": 0.4666666666666666,
443
+ "eval_dim_768_cosine_recall@1": 0.0811965811965812,
444
+ "eval_dim_768_cosine_recall@10": 0.47466422466422464,
445
+ "eval_dim_768_cosine_recall@3": 0.21978021978021975,
446
+ "eval_dim_768_cosine_recall@5": 0.2909035409035409,
447
+ "eval_runtime": 7.0685,
448
+ "eval_samples_per_second": 0.0,
449
+ "eval_sequential_score": 0.5232746482532176,
450
+ "eval_steps_per_second": 0.0,
451
+ "step": 21
452
+ },
453
+ {
454
+ "epoch": 3.1818181818181817,
455
+ "grad_norm": 214.07061767578125,
456
+ "learning_rate": 1.4383711467890776e-05,
457
+ "loss": 4.4446,
458
+ "step": 22
459
+ },
460
+ {
461
+ "epoch": 3.3636363636363638,
462
+ "grad_norm": 234.80874633789062,
463
+ "learning_rate": 1.3746065934159123e-05,
464
+ "loss": 5.7244,
465
+ "step": 23
466
+ },
467
+ {
468
+ "epoch": 3.5454545454545454,
469
+ "grad_norm": 337.2349853515625,
470
+ "learning_rate": 1.3090169943749475e-05,
471
+ "loss": 7.1394,
472
+ "step": 24
473
+ },
474
+ {
475
+ "epoch": 3.7272727272727275,
476
+ "grad_norm": 562.9892578125,
477
+ "learning_rate": 1.2419218955996677e-05,
478
+ "loss": 16.7583,
479
+ "step": 25
480
+ },
481
+ {
482
+ "epoch": 3.909090909090909,
483
+ "grad_norm": 1042.4781494140625,
484
+ "learning_rate": 1.1736481776669307e-05,
485
+ "loss": 11.3515,
486
+ "step": 26
487
+ },
488
+ {
489
+ "epoch": 4.0,
490
+ "grad_norm": 663.6896362304688,
491
+ "learning_rate": 1.1045284632676535e-05,
492
+ "loss": 8.813,
493
+ "step": 27
494
+ },
495
+ {
496
+ "epoch": 4.090909090909091,
497
+ "grad_norm": 799.2377319335938,
498
+ "learning_rate": 1.0348994967025012e-05,
499
+ "loss": 6.9124,
500
+ "step": 28
501
+ },
502
+ {
503
+ "epoch": 4.090909090909091,
504
+ "eval_dim_1024_cosine_accuracy@1": 0.47619047619047616,
505
+ "eval_dim_1024_cosine_accuracy@10": 0.5714285714285714,
506
+ "eval_dim_1024_cosine_accuracy@3": 0.47619047619047616,
507
+ "eval_dim_1024_cosine_accuracy@5": 0.47619047619047616,
508
+ "eval_dim_1024_cosine_map@100": 0.6149109740313521,
509
+ "eval_dim_1024_cosine_mrr@10": 0.49092970521541945,
510
+ "eval_dim_1024_cosine_ndcg@10": 0.5159384546892658,
511
+ "eval_dim_1024_cosine_precision@1": 0.47619047619047616,
512
+ "eval_dim_1024_cosine_precision@10": 0.4,
513
+ "eval_dim_1024_cosine_precision@3": 0.4603174603174603,
514
+ "eval_dim_1024_cosine_precision@5": 0.419047619047619,
515
+ "eval_dim_1024_cosine_recall@1": 0.07822039072039072,
516
+ "eval_dim_1024_cosine_recall@10": 0.4449023199023199,
517
+ "eval_dim_1024_cosine_recall@3": 0.21085164835164832,
518
+ "eval_dim_1024_cosine_recall@5": 0.27602258852258854,
519
+ "eval_dim_128_cosine_accuracy@1": 0.47619047619047616,
520
+ "eval_dim_128_cosine_accuracy@10": 0.5238095238095238,
521
+ "eval_dim_128_cosine_accuracy@3": 0.47619047619047616,
522
+ "eval_dim_128_cosine_accuracy@5": 0.47619047619047616,
523
+ "eval_dim_128_cosine_map@100": 0.6025310247157158,
524
+ "eval_dim_128_cosine_mrr@10": 0.48299319727891155,
525
+ "eval_dim_128_cosine_ndcg@10": 0.49922091065744895,
526
+ "eval_dim_128_cosine_precision@1": 0.47619047619047616,
527
+ "eval_dim_128_cosine_precision@10": 0.3761904761904762,
528
+ "eval_dim_128_cosine_precision@3": 0.4603174603174603,
529
+ "eval_dim_128_cosine_precision@5": 0.419047619047619,
530
+ "eval_dim_128_cosine_recall@1": 0.07822039072039072,
531
+ "eval_dim_128_cosine_recall@10": 0.42506105006105005,
532
+ "eval_dim_128_cosine_recall@3": 0.21085164835164832,
533
+ "eval_dim_128_cosine_recall@5": 0.27602258852258854,
534
+ "eval_dim_256_cosine_accuracy@1": 0.5238095238095238,
535
+ "eval_dim_256_cosine_accuracy@10": 0.5714285714285714,
536
+ "eval_dim_256_cosine_accuracy@3": 0.5238095238095238,
537
+ "eval_dim_256_cosine_accuracy@5": 0.5238095238095238,
538
+ "eval_dim_256_cosine_map@100": 0.6411393184007045,
539
+ "eval_dim_256_cosine_mrr@10": 0.5306122448979591,
540
+ "eval_dim_256_cosine_ndcg@10": 0.5468399582764966,
541
+ "eval_dim_256_cosine_precision@1": 0.5238095238095238,
542
+ "eval_dim_256_cosine_precision@10": 0.4238095238095239,
543
+ "eval_dim_256_cosine_precision@3": 0.5079365079365079,
544
+ "eval_dim_256_cosine_precision@5": 0.4666666666666666,
545
+ "eval_dim_256_cosine_recall@1": 0.08005189255189255,
546
+ "eval_dim_256_cosine_recall@10": 0.4433760683760684,
547
+ "eval_dim_256_cosine_recall@3": 0.21634615384615385,
548
+ "eval_dim_256_cosine_recall@5": 0.28518009768009767,
549
+ "eval_dim_512_cosine_accuracy@1": 0.47619047619047616,
550
+ "eval_dim_512_cosine_accuracy@10": 0.5238095238095238,
551
+ "eval_dim_512_cosine_accuracy@3": 0.47619047619047616,
552
+ "eval_dim_512_cosine_accuracy@5": 0.47619047619047616,
553
+ "eval_dim_512_cosine_map@100": 0.5978106306698094,
554
+ "eval_dim_512_cosine_mrr@10": 0.48299319727891155,
555
+ "eval_dim_512_cosine_ndcg@10": 0.49922091065744895,
556
+ "eval_dim_512_cosine_precision@1": 0.47619047619047616,
557
+ "eval_dim_512_cosine_precision@10": 0.3761904761904762,
558
+ "eval_dim_512_cosine_precision@3": 0.4603174603174603,
559
+ "eval_dim_512_cosine_precision@5": 0.419047619047619,
560
+ "eval_dim_512_cosine_recall@1": 0.07822039072039072,
561
+ "eval_dim_512_cosine_recall@10": 0.42506105006105005,
562
+ "eval_dim_512_cosine_recall@3": 0.21085164835164832,
563
+ "eval_dim_512_cosine_recall@5": 0.27602258852258854,
564
+ "eval_dim_64_cosine_accuracy@1": 0.47619047619047616,
565
+ "eval_dim_64_cosine_accuracy@10": 0.5238095238095238,
566
+ "eval_dim_64_cosine_accuracy@3": 0.47619047619047616,
567
+ "eval_dim_64_cosine_accuracy@5": 0.47619047619047616,
568
+ "eval_dim_64_cosine_map@100": 0.5960251374266525,
569
+ "eval_dim_64_cosine_mrr@10": 0.48299319727891155,
570
+ "eval_dim_64_cosine_ndcg@10": 0.49922091065744895,
571
+ "eval_dim_64_cosine_precision@1": 0.47619047619047616,
572
+ "eval_dim_64_cosine_precision@10": 0.3761904761904762,
573
+ "eval_dim_64_cosine_precision@3": 0.4603174603174603,
574
+ "eval_dim_64_cosine_precision@5": 0.419047619047619,
575
+ "eval_dim_64_cosine_recall@1": 0.07822039072039072,
576
+ "eval_dim_64_cosine_recall@10": 0.42506105006105005,
577
+ "eval_dim_64_cosine_recall@3": 0.21085164835164832,
578
+ "eval_dim_64_cosine_recall@5": 0.27602258852258854,
579
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
580
+ "eval_dim_768_cosine_accuracy@10": 0.5714285714285714,
581
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
582
+ "eval_dim_768_cosine_accuracy@5": 0.5238095238095238,
583
+ "eval_dim_768_cosine_map@100": 0.6351788392177582,
584
+ "eval_dim_768_cosine_mrr@10": 0.5306122448979591,
585
+ "eval_dim_768_cosine_ndcg@10": 0.5468399582764966,
586
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
587
+ "eval_dim_768_cosine_precision@10": 0.4238095238095238,
588
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
589
+ "eval_dim_768_cosine_precision@5": 0.4666666666666666,
590
+ "eval_dim_768_cosine_recall@1": 0.08218864468864469,
591
+ "eval_dim_768_cosine_recall@10": 0.46474358974358976,
592
+ "eval_dim_768_cosine_recall@3": 0.22275641025641024,
593
+ "eval_dim_768_cosine_recall@5": 0.2958638583638584,
594
+ "eval_runtime": 7.0378,
595
+ "eval_samples_per_second": 0.0,
596
+ "eval_sequential_score": 0.49922091065744895,
597
+ "eval_steps_per_second": 0.0,
598
+ "step": 28
599
+ },
600
+ {
601
+ "epoch": 4.181818181818182,
602
+ "grad_norm": 278.7558898925781,
603
+ "learning_rate": 9.651005032974994e-06,
604
+ "loss": 6.1814,
605
+ "step": 29
606
+ },
607
+ {
608
+ "epoch": 4.363636363636363,
609
+ "grad_norm": 312.5499572753906,
610
+ "learning_rate": 8.954715367323468e-06,
611
+ "loss": 7.1606,
612
+ "step": 30
613
+ },
614
+ {
615
+ "epoch": 4.545454545454545,
616
+ "grad_norm": 164.37811279296875,
617
+ "learning_rate": 8.263518223330698e-06,
618
+ "loss": 5.0888,
619
+ "step": 31
620
+ },
621
+ {
622
+ "epoch": 4.7272727272727275,
623
+ "grad_norm": 288.98602294921875,
624
+ "learning_rate": 7.580781044003324e-06,
625
+ "loss": 5.0684,
626
+ "step": 32
627
+ },
628
+ {
629
+ "epoch": 4.909090909090909,
630
+ "grad_norm": 384.2759704589844,
631
+ "learning_rate": 6.909830056250527e-06,
632
+ "loss": 6.7382,
633
+ "step": 33
634
+ },
635
+ {
636
+ "epoch": 5.0,
637
+ "grad_norm": 906.340576171875,
638
+ "learning_rate": 6.25393406584088e-06,
639
+ "loss": 7.0497,
640
+ "step": 34
641
+ },
642
+ {
643
+ "epoch": 5.090909090909091,
644
+ "grad_norm": 403.593017578125,
645
+ "learning_rate": 5.616288532109225e-06,
646
+ "loss": 6.582,
647
+ "step": 35
648
+ },
649
+ {
650
+ "epoch": 5.090909090909091,
651
+ "eval_dim_1024_cosine_accuracy@1": 0.5238095238095238,
652
+ "eval_dim_1024_cosine_accuracy@10": 0.6190476190476191,
653
+ "eval_dim_1024_cosine_accuracy@3": 0.5238095238095238,
654
+ "eval_dim_1024_cosine_accuracy@5": 0.5238095238095238,
655
+ "eval_dim_1024_cosine_map@100": 0.6534286699882501,
656
+ "eval_dim_1024_cosine_mrr@10": 0.5374149659863945,
657
+ "eval_dim_1024_cosine_ndcg@10": 0.5598242514045669,
658
+ "eval_dim_1024_cosine_precision@1": 0.5238095238095238,
659
+ "eval_dim_1024_cosine_precision@10": 0.4428571428571429,
660
+ "eval_dim_1024_cosine_precision@3": 0.5079365079365079,
661
+ "eval_dim_1024_cosine_precision@5": 0.4666666666666666,
662
+ "eval_dim_1024_cosine_recall@1": 0.08218864468864469,
663
+ "eval_dim_1024_cosine_recall@10": 0.4766483516483517,
664
+ "eval_dim_1024_cosine_recall@3": 0.22275641025641024,
665
+ "eval_dim_1024_cosine_recall@5": 0.2958638583638584,
666
+ "eval_dim_128_cosine_accuracy@1": 0.5238095238095238,
667
+ "eval_dim_128_cosine_accuracy@10": 0.6190476190476191,
668
+ "eval_dim_128_cosine_accuracy@3": 0.5238095238095238,
669
+ "eval_dim_128_cosine_accuracy@5": 0.5238095238095238,
670
+ "eval_dim_128_cosine_map@100": 0.6478872365910466,
671
+ "eval_dim_128_cosine_mrr@10": 0.5374149659863945,
672
+ "eval_dim_128_cosine_ndcg@10": 0.5598242514045669,
673
+ "eval_dim_128_cosine_precision@1": 0.5238095238095238,
674
+ "eval_dim_128_cosine_precision@10": 0.4428571428571429,
675
+ "eval_dim_128_cosine_precision@3": 0.5079365079365079,
676
+ "eval_dim_128_cosine_precision@5": 0.4666666666666666,
677
+ "eval_dim_128_cosine_recall@1": 0.0811965811965812,
678
+ "eval_dim_128_cosine_recall@10": 0.46672771672771673,
679
+ "eval_dim_128_cosine_recall@3": 0.21978021978021975,
680
+ "eval_dim_128_cosine_recall@5": 0.2909035409035409,
681
+ "eval_dim_256_cosine_accuracy@1": 0.6190476190476191,
682
+ "eval_dim_256_cosine_accuracy@10": 0.6666666666666666,
683
+ "eval_dim_256_cosine_accuracy@3": 0.6190476190476191,
684
+ "eval_dim_256_cosine_accuracy@5": 0.6190476190476191,
685
+ "eval_dim_256_cosine_map@100": 0.6975707466438095,
686
+ "eval_dim_256_cosine_mrr@10": 0.6258503401360545,
687
+ "eval_dim_256_cosine_ndcg@10": 0.6420780535145918,
688
+ "eval_dim_256_cosine_precision@1": 0.6190476190476191,
689
+ "eval_dim_256_cosine_precision@10": 0.5190476190476192,
690
+ "eval_dim_256_cosine_precision@3": 0.6031746031746031,
691
+ "eval_dim_256_cosine_precision@5": 0.5619047619047619,
692
+ "eval_dim_256_cosine_recall@1": 0.08600427350427349,
693
+ "eval_dim_256_cosine_recall@10": 0.5028998778998779,
694
+ "eval_dim_256_cosine_recall@3": 0.2342032967032967,
695
+ "eval_dim_256_cosine_recall@5": 0.31494200244200243,
696
+ "eval_dim_512_cosine_accuracy@1": 0.5238095238095238,
697
+ "eval_dim_512_cosine_accuracy@10": 0.6190476190476191,
698
+ "eval_dim_512_cosine_accuracy@3": 0.5238095238095238,
699
+ "eval_dim_512_cosine_accuracy@5": 0.5238095238095238,
700
+ "eval_dim_512_cosine_map@100": 0.6492208787775379,
701
+ "eval_dim_512_cosine_mrr@10": 0.5374149659863945,
702
+ "eval_dim_512_cosine_ndcg@10": 0.5598242514045669,
703
+ "eval_dim_512_cosine_precision@1": 0.5238095238095238,
704
+ "eval_dim_512_cosine_precision@10": 0.4428571428571429,
705
+ "eval_dim_512_cosine_precision@3": 0.5079365079365079,
706
+ "eval_dim_512_cosine_precision@5": 0.4666666666666666,
707
+ "eval_dim_512_cosine_recall@1": 0.08218864468864469,
708
+ "eval_dim_512_cosine_recall@10": 0.4766483516483517,
709
+ "eval_dim_512_cosine_recall@3": 0.22275641025641024,
710
+ "eval_dim_512_cosine_recall@5": 0.2958638583638584,
711
+ "eval_dim_64_cosine_accuracy@1": 0.42857142857142855,
712
+ "eval_dim_64_cosine_accuracy@10": 0.5714285714285714,
713
+ "eval_dim_64_cosine_accuracy@3": 0.47619047619047616,
714
+ "eval_dim_64_cosine_accuracy@5": 0.47619047619047616,
715
+ "eval_dim_64_cosine_map@100": 0.5872011588310861,
716
+ "eval_dim_64_cosine_mrr@10": 0.45804988662131524,
717
+ "eval_dim_64_cosine_ndcg@10": 0.4917595713548203,
718
+ "eval_dim_64_cosine_precision@1": 0.42857142857142855,
719
+ "eval_dim_64_cosine_precision@10": 0.3952380952380953,
720
+ "eval_dim_64_cosine_precision@3": 0.4444444444444445,
721
+ "eval_dim_64_cosine_precision@5": 0.419047619047619,
722
+ "eval_dim_64_cosine_recall@1": 0.054410866910866905,
723
+ "eval_dim_64_cosine_recall@10": 0.43696581196581197,
724
+ "eval_dim_64_cosine_recall@3": 0.18704212454212454,
725
+ "eval_dim_64_cosine_recall@5": 0.27602258852258854,
726
+ "eval_dim_768_cosine_accuracy@1": 0.5238095238095238,
727
+ "eval_dim_768_cosine_accuracy@10": 0.6190476190476191,
728
+ "eval_dim_768_cosine_accuracy@3": 0.5238095238095238,
729
+ "eval_dim_768_cosine_accuracy@5": 0.5238095238095238,
730
+ "eval_dim_768_cosine_map@100": 0.653075337994289,
731
+ "eval_dim_768_cosine_mrr@10": 0.5374149659863945,
732
+ "eval_dim_768_cosine_ndcg@10": 0.5598242514045669,
733
+ "eval_dim_768_cosine_precision@1": 0.5238095238095238,
734
+ "eval_dim_768_cosine_precision@10": 0.4428571428571429,
735
+ "eval_dim_768_cosine_precision@3": 0.5079365079365079,
736
+ "eval_dim_768_cosine_precision@5": 0.4666666666666666,
737
+ "eval_dim_768_cosine_recall@1": 0.08218864468864469,
738
+ "eval_dim_768_cosine_recall@10": 0.4766483516483517,
739
+ "eval_dim_768_cosine_recall@3": 0.22275641025641024,
740
+ "eval_dim_768_cosine_recall@5": 0.2958638583638584,
741
+ "eval_runtime": 7.0134,
742
+ "eval_samples_per_second": 0.0,
743
+ "eval_sequential_score": 0.4917595713548203,
744
+ "eval_steps_per_second": 0.0,
745
+ "step": 35
746
+ }
747
+ ],
748
+ "logging_steps": 1,
749
+ "max_steps": 50,
750
+ "num_input_tokens_seen": 0,
751
+ "num_train_epochs": 10,
752
+ "save_steps": 500,
753
+ "stateful_callbacks": {
754
+ "EarlyStoppingCallback": {
755
+ "args": {
756
+ "early_stopping_patience": 3,
757
+ "early_stopping_threshold": 0.0
758
+ },
759
+ "attributes": {
760
+ "early_stopping_patience_counter": 3
761
+ }
762
+ },
763
+ "TrainerControl": {
764
+ "args": {
765
+ "should_epoch_stop": false,
766
+ "should_evaluate": false,
767
+ "should_log": false,
768
+ "should_save": true,
769
+ "should_training_stop": true
770
+ },
771
+ "attributes": {}
772
+ }
773
+ },
774
+ "total_flos": 0.0,
775
+ "train_batch_size": 8,
776
+ "trial_name": null,
777
+ "trial_params": null
778
+ }
checkpoint-35/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14debc6c3f8c5edee5db8d97a3a78a007d313a13e4b96f43026da543b59bef8c
3
+ size 6097