{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":144993086,"defaultBranch":"master","name":"NeuroX","ownerLogin":"fdalvi","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2018-08-16T13:41:02.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/859719?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1693384229.0","currentOid":""},"activityList":{"items":[{"before":"8b8e0d967cc4d0030527947c34e78c194daa850a","after":"c8276fa2e587d0f2b3207ede25788074ecb117ed","ref":"refs/heads/feat/transformers_seq2seq_extraction","pushedAt":"2023-10-04T13:09:44.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Add preliminary implementation for decoder extraction","shortMessageHtmlLink":"Add preliminary implementation for decoder extraction"}},{"before":"0e9a9e1b8d13b16de2977b81661eab1df8641f47","after":"8b8e0d967cc4d0030527947c34e78c194daa850a","ref":"refs/heads/feat/transformers_seq2seq_extraction","pushedAt":"2023-10-04T12:49:59.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Add preliminary implementation for decoder extraction","shortMessageHtmlLink":"Add preliminary implementation for decoder extraction"}},{"before":"7434f60208c773a70a94240ef2a162c39703f5b4","after":"0e9a9e1b8d13b16de2977b81661eab1df8641f47","ref":"refs/heads/feat/transformers_seq2seq_extraction","pushedAt":"2023-10-02T07:15:39.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Use a token from the vocab and make sure its unbroken after tokenizer for context indicator","shortMessageHtmlLink":"Use a token from the vocab and make sure its unbroken after tokenizer…"}},{"before":"558ff67a492613eb4d8eb549062e5efe3a336b23","after":"8119c575fc1c861de3fa49004ecffabb0a292de0","ref":"refs/heads/tests/unit-tests","pushedAt":"2023-09-01T16:35:07.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"modified probeless unit tests","shortMessageHtmlLink":"modified probeless unit tests"}},{"before":"2f05384d7ca1eb2bd27f9d4d60b5cc7626bf97de","after":"558ff67a492613eb4d8eb549062e5efe3a336b23","ref":"refs/heads/tests/unit-tests","pushedAt":"2023-09-01T14:01:53.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"added prelimenary unit tests for probeless","shortMessageHtmlLink":"added prelimenary unit tests for probeless"}},{"before":"434f9a6005c287dee1f166f5d8bf486d545b8e5a","after":"6c47effd07a561515a587fafca456f94385fa775","ref":"refs/heads/tests/regression-tests","pushedAt":"2023-08-31T16:09:46.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"formatted the second version of the ranking","shortMessageHtmlLink":"formatted the second version of the ranking"}},{"before":"e67395dd9dfa0096b80ef53d4503ae6e35bc8f7e","after":"434f9a6005c287dee1f166f5d8bf486d545b8e5a","ref":"refs/heads/tests/regression-tests","pushedAt":"2023-08-30T19:54:20.000Z","pushType":"push","commitsCount":4,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"formatted code","shortMessageHtmlLink":"formatted code"}},{"before":null,"after":"b35a42a3aec67eac94706c71510dcba4d8dca031","ref":"refs/heads/activation-extraction","pushedAt":"2023-08-30T08:30:29.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"Added verbose option for activation extraction","shortMessageHtmlLink":"Added verbose option for activation extraction"}},{"before":"b35a42a3aec67eac94706c71510dcba4d8dca031","after":"a9ac7cd90be89d72c5ed55d963e2843f039303fc","ref":"refs/heads/master","pushedAt":"2023-08-30T08:17:10.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Implement special tokens extraction for transformers (#41)\n\nThe transformers extractor did not support extraction of representations\r\nfor special tokens like `[CLS]`, `[SEP]`, `[PAD]` etc. This commit adds\r\nan additional parameter to also extract these along with the original\r\ntokens.\r\n\r\nDetailed commit history:\r\n\r\n* Implement special tokens extraction for transformers\r\n\r\nThe transformers extractor did not support extraction of representations\r\nfor special tokens like `[CLS]`, `[SEP]`, `[PAD]` etc. This commit adds\r\nan additional parameter to also extract these along with the original\r\ntokens.\r\n\r\n* Add check for special/dropped token adjacency\r\n\r\nWhen a special token follows a token that was completely dropped by\r\nthe tokenizer, or vice-versa, we do not have enough information to\r\nrebuild the original input and now explictly raise an exception.\r\n\r\n* Refactor transformers tests for clarity\r\n\r\n* Add tests for long inputs\r\n\r\n* Add command line option in extractor\r\n\r\n* Format code\r\n\r\n* Add documentation for extraction functions","shortMessageHtmlLink":"Implement special tokens extraction for transformers (#41)"}},{"before":"a9ac7cd90be89d72c5ed55d963e2843f039303fc","after":"b35a42a3aec67eac94706c71510dcba4d8dca031","ref":"refs/heads/master","pushedAt":"2023-08-30T08:06:12.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"Added verbose option for activation extraction","shortMessageHtmlLink":"Added verbose option for activation extraction"}},{"before":null,"after":"2f05384d7ca1eb2bd27f9d4d60b5cc7626bf97de","ref":"refs/heads/tests/unit-tests","pushedAt":"2023-08-30T07:42:00.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"added prelimenary unit test for probeless","shortMessageHtmlLink":"added prelimenary unit test for probeless"}},{"before":null,"after":"e67395dd9dfa0096b80ef53d4503ae6e35bc8f7e","ref":"refs/heads/tests/regression-tests","pushedAt":"2023-08-30T07:11:50.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"modified format","shortMessageHtmlLink":"modified format"}},{"before":null,"after":"19c8d15fe91304be0e1867e37b74f38309983d77","ref":"refs/heads/feat/activation-extraction","pushedAt":"2023-08-21T14:59:23.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"Added verbose option for activation extraction","shortMessageHtmlLink":"Added verbose option for activation extraction"}},{"before":null,"after":"9356c3477cde9611c90c719bfcd236c24134093e","ref":"refs/heads/notebook-tutorials","pushedAt":"2023-07-06T14:10:23.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"baselmousi","name":"Basel Mousi","path":"/baselmousi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/59998313?s=80&v=4"},"commit":{"message":"Added data for tutorial","shortMessageHtmlLink":"Added data for tutorial"}},{"before":null,"after":"919031132cef23c24cfc08ffa6053e649dea9939","ref":"refs/heads/feat/transformers_multi_sentence_extraction","pushedAt":"2023-05-25T14:21:05.300Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Preliminary implemention of multi-sentence support","shortMessageHtmlLink":"Preliminary implemention of multi-sentence support"}},{"before":null,"after":"7434f60208c773a70a94240ef2a162c39703f5b4","ref":"refs/heads/feat/transformers_seq2seq_extraction","pushedAt":"2023-05-14T13:06:54.739Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Proof of Concept for Seq2Seq model extraction\n\nOnly encoder extraction is supported currently. Structure very likely\nto change (probably with a separate module for seq2seq models).","shortMessageHtmlLink":"Proof of Concept for Seq2Seq model extraction"}},{"before":"c553671deb90ad88db66f38384c6086022727dfb","after":"a9ac7cd90be89d72c5ed55d963e2843f039303fc","ref":"refs/heads/master","pushedAt":"2023-04-12T07:14:13.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Implement special tokens extraction for transformers (#41)\n\nThe transformers extractor did not support extraction of representations\r\nfor special tokens like `[CLS]`, `[SEP]`, `[PAD]` etc. This commit adds\r\nan additional parameter to also extract these along with the original\r\ntokens.\r\n\r\nDetailed commit history:\r\n\r\n* Implement special tokens extraction for transformers\r\n\r\nThe transformers extractor did not support extraction of representations\r\nfor special tokens like `[CLS]`, `[SEP]`, `[PAD]` etc. This commit adds\r\nan additional parameter to also extract these along with the original\r\ntokens.\r\n\r\n* Add check for special/dropped token adjacency\r\n\r\nWhen a special token follows a token that was completely dropped by\r\nthe tokenizer, or vice-versa, we do not have enough information to\r\nrebuild the original input and now explictly raise an exception.\r\n\r\n* Refactor transformers tests for clarity\r\n\r\n* Add tests for long inputs\r\n\r\n* Add command line option in extractor\r\n\r\n* Format code\r\n\r\n* Add documentation for extraction functions","shortMessageHtmlLink":"Implement special tokens extraction for transformers (#41)"}},{"before":"9a4b3bdcaa1bdc203aad24d598e8af131bd4c7f6","after":"46610c8c27617727b2eb02b2f0ec05e6c8084382","ref":"refs/heads/feat/transformers_special_tokens_extraction","pushedAt":"2023-04-12T07:06:30.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Add documentation for extraction functions","shortMessageHtmlLink":"Add documentation for extraction functions"}},{"before":"6e0d5eb8e95d8be460621b70aca9d2f71dcc8ada","after":"9a4b3bdcaa1bdc203aad24d598e8af131bd4c7f6","ref":"refs/heads/feat/transformers_special_tokens_extraction","pushedAt":"2023-04-03T09:25:43.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Format code","shortMessageHtmlLink":"Format code"}},{"before":"155bfdd02de45e0f5faa09cff6ac82bc7a93bfbf","after":"6e0d5eb8e95d8be460621b70aca9d2f71dcc8ada","ref":"refs/heads/feat/transformers_special_tokens_extraction","pushedAt":"2023-04-03T09:22:41.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Add command line option in extractor","shortMessageHtmlLink":"Add command line option in extractor"}},{"before":"06e2dfb828244a816aa617836d30f5cf671e2b65","after":"155bfdd02de45e0f5faa09cff6ac82bc7a93bfbf","ref":"refs/heads/feat/transformers_special_tokens_extraction","pushedAt":"2023-03-29T13:44:50.137Z","pushType":"push","commitsCount":2,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Add tests for long inputs","shortMessageHtmlLink":"Add tests for long inputs"}},{"before":"e671251a53f2eb551bce02b2336d23ed8625955f","after":"06e2dfb828244a816aa617836d30f5cf671e2b65","ref":"refs/heads/feat/transformers_special_tokens_extraction","pushedAt":"2023-03-29T11:47:25.808Z","pushType":"push","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Add check for special/dropped token adjacency\n\nWhen a special token follows a token that was completely dropped by\nthe tokenizer, or vice-versa, we do not have enough information to\nrebuild the original input and now explictly raise an exception.","shortMessageHtmlLink":"Add check for special/dropped token adjacency"}},{"before":null,"after":"e671251a53f2eb551bce02b2336d23ed8625955f","ref":"refs/heads/feat/transformers_special_tokens_extraction","pushedAt":"2023-03-29T05:05:33.233Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Implement special tokens extraction for transformers\n\nThe transformers extractor did not support extraction of representations\nfor special tokens like `[CLS]`, `[SEP]`, `[PAD]` etc. This commit adds\nan additional parameter to also extract these along with the original\ntokens.","shortMessageHtmlLink":"Implement special tokens extraction for transformers"}},{"before":"32953971783453b5454e7c8b27ca038c9072a93d","after":"c553671deb90ad88db66f38384c6086022727dfb","ref":"refs/heads/master","pushedAt":"2023-03-28T09:41:13.952Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Bugfix for varying tokenizations in transformers (#40)\n\nSome transformers tokenizers split the _same_ token into different\r\nsubwords depending on whether the token appears at the beginning,\r\nmiddle or end of the sentence. The detokenization algorithm in the\r\nextraction code now accounts for such cases.\r\n\r\nSample case:\r\n```\r\npython -m neurox.data.extraction.transformers_extractor flaubert/flaubert_base_cased <(echo \"qu' les commotions cérébrales sont devenu si courantes dans ce sport qu' on les considére presque comme la routine qu'\") test.json\r\n```","shortMessageHtmlLink":"Bugfix for varying tokenizations in transformers (#40)"}},{"before":null,"after":"57c65c0abab45971cb5b1dd3984174d3f18344a1","ref":"refs/heads/fix/transformers_detokenization","pushedAt":"2023-03-28T09:38:30.369Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Bugfix for varying tokenizations in transformers\n\nSome transformers tokenizers split the _same_ token into different\nsubwords depending on whether the token appears at the beginning,\nmiddle or end of the sentence. The detokenization algorithm in the\nextraction code now accounts for such cases.\n\nSample case:\n```\npython -m neurox.data.extraction.transformers_extractor flaubert/flaubert_base_cased <(echo \"qu' les commotions cérébrales sont devenu si courantes dans ce sport qu' on les considére presque comme la routine qu'\") test.json\n```","shortMessageHtmlLink":"Bugfix for varying tokenizations in transformers"}},{"before":"964d269adc9cc35d8483e17db3a816360700a7cf","after":"32953971783453b5454e7c8b27ca038c9072a93d","ref":"refs/heads/master","pushedAt":"2023-03-28T07:04:54.153Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Fix Deprecated warnings in annotate.py (#38)\n\nannotate.py was using collections.Callable in a deprecated fashion,\r\nand a couple comments were not explicitly marked as raw strings and\r\nled to parsing warnings. These have now been fixed.","shortMessageHtmlLink":"Fix Deprecated warnings in annotate.py (#38)"}},{"before":"6237948d3dc52aeb3bf341a17c1bdf78c91bfb83","after":"b98213585f2ff0fd208aa87694027dddfba159df","ref":"refs/heads/fix/annotate_warnings","pushedAt":"2023-03-28T07:02:24.682Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Fix Deprecated warnings in annotate.py\n\nannotate.py was using collections.Callable in a deprecated fashion,\nand a couple comments were not explicitly marked as raw strings and\nled to parsing warnings. These have now been fixed.","shortMessageHtmlLink":"Fix Deprecated warnings in annotate.py"}},{"before":"f2314ccd964f744eb3209f1438fa53704a6665b2","after":"964d269adc9cc35d8483e17db3a816360700a7cf","ref":"refs/heads/master","pushedAt":"2023-03-28T06:58:16.991Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"GH actions: Fixes for Python 3.6 and `ufmt` changes (#39)\n\n* GH action: Fix for Python 3.6\r\n\r\n* Manual `ufmt` fixes that only show up on linux\r\n\r\n`ufmt` passes on macOS, but fails on linux for some files. This\r\nfixes this so that it passes on both.","shortMessageHtmlLink":"GH actions: Fixes for Python 3.6 and ufmt
changes (#39)"}},{"before":"d93646ece1cb095b814b384b85021339c1872ee4","after":"eac6b4bf6ac3c6a88d67d17fad62746498f0f568","ref":"refs/heads/fix/gh_workflows","pushedAt":"2023-03-28T06:55:06.034Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Manual `ufmt` fixes that only show up on linux\n\n`ufmt` passes on macOS, but fails on linux for some files. This\nfixes this so that it passes on both.","shortMessageHtmlLink":"Manual ufmt
fixes that only show up on linux"}},{"before":"621e1b0afbeb11714c965ac20c6dbc89fab459f3","after":"d93646ece1cb095b814b384b85021339c1872ee4","ref":"refs/heads/fix/gh_workflows","pushedAt":"2023-03-28T06:50:29.037Z","pushType":"push","commitsCount":1,"pusher":{"login":"fdalvi","name":"Fahim Dalvi","path":"/fdalvi","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/859719?s=80&v=4"},"commit":{"message":"Manual `ufmt` fixes that only show up on linux\n\n`ufmt` passes on macOS, but fails on linux for some files. This\nfixes this so that it passes on both.","shortMessageHtmlLink":"Manual ufmt
fixes that only show up on linux"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAADj0BGugA","startCursor":null,"endCursor":null}},"title":"Activity · fdalvi/NeuroX"}