Scheduled Commit

data/retrieval_individual-8d38bb89-1e1a-471a-8b9e-35c1f784690e.jsonl

@@ -68,3 +68,5 @@
 {"tstamp": 1722365480.6099, "task_type": "retrieval", "type": "chat", "model": "Salesforce/SFR-Embedding-2_R", "gen_params": {}, "start": 1722365480.2948, "finish": 1722365480.6099, "ip": "", "conv_id": "960bea43fd02452d9429b486a6719b0e", "model_name": "Salesforce/SFR-Embedding-2_R", "prompt": "what is dense retrieval", "output": [["what is dense retrieval", "Information retrieval\n\nInformation retrieval (IR) in computing and information science is the task of identifying and retrieving information system resources that are relevant to an information need.  The information need can be specified in the form of a search query. In the case of document retrieval, queries can be based on full-text or other content-based indexing. Information retrieval is the science of searching for information in a document, searching for documents themselves, and also searching for the metadata that describes data, and for databases of texts, images or sounds.\nAutomated information retrieval systems are used to reduce what has been called information overload. An IR system is a software system that provides access to books, journals and other documents; it also stores and manages those documents. Web search engines are the most visible IR applications."]], "corpus": "wikipedia"}
 {"tstamp": 1722366414.2932, "task_type": "retrieval", "type": "chat", "model": "embed-english-v3.0", "gen_params": {}, "start": 1722366413.3732, "finish": 1722366414.2932, "ip": "", "conv_id": "3406771838ae4294a308437176647a93", "model_name": "embed-english-v3.0", "prompt": "Research introducing a technique where visual vocabulary learned from image tags is used for generating captions for novel objects.", "output": [["Research introducing a technique where visual vocabulary learned from image tags is used for generating captions for novel objects.", "Title: Captioning Images with Novel Objects via Online Vocabulary Expansion\n\nAbstract: In this study, we introduce a low cost method for generating descriptions from images containing novel objects. Generally, constructing a model, which can explain images with novel objects, is costly because of the following: (1) collecting a large amount of data for each category, and (2) retraining the entire system. If humans see a small number of novel objects, they are able to estimate their properties by associating their appearance with known objects. Accordingly, we propose a method that can explain images with novel objects without retraining using the word embeddings of the objects estimated from only a small number of image features of the objects. The method can be integrated with general image-captioning models. The experimental results show the effectiveness of our approach."]], "corpus": "arxiv"}
 {"tstamp": 1722366414.2932, "task_type": "retrieval", "type": "chat", "model": "nomic-ai/nomic-embed-text-v1.5", "gen_params": {}, "start": 1722366413.3732, "finish": 1722366414.2932, "ip": "", "conv_id": "3f5d842f29464f66b1432b7a355fab52", "model_name": "nomic-ai/nomic-embed-text-v1.5", "prompt": "Research introducing a technique where visual vocabulary learned from image tags is used for generating captions for novel objects.", "output": [["Research introducing a technique where visual vocabulary learned from image tags is used for generating captions for novel objects.", "Title: Young massive star clusters in the era of the Hubble Space Telescope\n\nAbstract: The Hubble Space Telescope (HST) has been instrumental in the discovery of large numbers of extragalactic young massive star clusters (YMCs), often assumed to be proto-globular clusters (GCs). As a consequence, the field of YMC formation and evolution is thriving, generating major breakthroughs as well as controversies on annual (or shorter) time-scales. Here, I review the long-term survival chances of YMCs, hallmarks of intense starburst episodes often associated with violent galaxy interactions. In the absence of significant external perturbations, the key factor determining a cluster's long-term survival chances is the shape of its stellar initial mass function (IMF). It is, however, not straightforward to assess the IMF shape in unresolved extragalactic YMCs. I also discuss the latest progress in worldwide efforts to better understand the evolution of entire cluster populations, predominantly based on HST observations, and conclude that there is an increasing body of evidence that GC formation appears to be continuing until today; their long-term evolution crucially depends on their environmental conditions, however."]], "corpus": "arxiv"}
+{"tstamp": 1722366808.573, "task_type": "retrieval", "type": "chat", "model": "GritLM/GritLM-7B", "gen_params": {}, "start": 1722366806.2538, "finish": 1722366808.573, "ip": "", "conv_id": "8b7dbc40208847bdbcdaba8aafdeee5a", "model_name": "GritLM/GritLM-7B", "prompt": "hi how are you ", "output": [["hi how are you ", "Hi, How Are You\n\nHi, How Are You: The Unfinished Album is the sixth self-released music cassette album by singer-songwriter Daniel Johnston, recorded in September 1983. The subtitle refers to the length of the album; Johnston had intended for it to be an hour-long release, as a majority of his earlier releases were.\nBackground"]], "corpus": "wikipedia"}
+{"tstamp": 1722366808.573, "task_type": "retrieval", "type": "chat", "model": "Salesforce/SFR-Embedding-2_R", "gen_params": {}, "start": 1722366806.2538, "finish": 1722366808.573, "ip": "", "conv_id": "411e98668a6e4744bfdb95ba014a56c8", "model_name": "Salesforce/SFR-Embedding-2_R", "prompt": "hi how are you ", "output": [["hi how are you ", "Howdy\n\nHowdy is an informal salutation in the English language often used in the Southern United States. Originally a shortened form of the inquiry How do ye?, it was first used in Southern England in the 18th century.\nEtymology\nThe word derives from the phrase \"How do ye?,\" which was used in late 16th century England to ask about others' health. The phrase saw increasing degrees of dialectal contraction over time, first being contracted to \"how dee,\" then to \"howdy'ee,\" and finally to its modern form. Despite the etymology of the term, its modern usage is as a greeting and not as an inquiry.\nUse in different states\nIn many rural Southern and Western states, especially in Arizona, California, Colorado, Idaho, Montana, Nevada, New Mexico, Oregon, Texas, and Wyoming, howdy is commonly used in casual contexts as a standard greeting. The term is commonly associated in popular culture specifically with the state of Texas, and the usage of the term is a significant and recognizable component of Texan English. Howdy is also the official greeting of Texas A&M University, though the two are not always automatically associated."]], "corpus": "wikipedia"}