LateOn-regularized / special_tokens_map.json

LateOn-unsupervised + MUVERA projection regularization (lambda=0.15, center=both). Trained to produce embeddings that survive MUVERA FDE compression. Beats baseline on MUVERA retrieval: scifact +3%, nfcorpus +5%.

7ae6c85 verified about 2 months ago

Raw

History Blame Contribute Delete

581 Bytes

	{
	"cls_token": {
	"content": "[CLS]",
	"lstrip": false,
	"normalized": false,
	"rstrip": false,
	"single_word": false
	},
	"mask_token": {
	"content": "[MASK]",
	"lstrip": true,
	"normalized": false,
	"rstrip": false,
	"single_word": false
	},
	"pad_token": "[MASK]",
	"sep_token": {
	"content": "[SEP]",
	"lstrip": false,
	"normalized": false,
	"rstrip": false,
	"single_word": false
	},
	"unk_token": {
	"content": "[UNK]",
	"lstrip": false,
	"normalized": false,
	"rstrip": false,
	"single_word": false
	}
	}