{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2021:5H7BMODOI5NN2RW5CGOVG4ACJ4","short_pith_number":"pith:5H7BMODO","canonical_record":{"source":{"id":"2104.06599","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-04-14T02:56:14Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"278a8260ad95ea1ce912af58cf13afff65c148cfde31ed3d6e26bbcbc3292f43","abstract_canon_sha256":"5a0528f100109f5df7b47d76aa10a4157b0f23ec375c6afee9a3877580158731"},"schema_version":"1.0"},"canonical_sha256":"e9fe16386e475add46dd119d5370024f11ffdb5e0c7fb3c207e540630ebecbdf","source":{"kind":"arxiv","id":"2104.06599","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2104.06599","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"arxiv_version","alias_value":"2104.06599v1","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2104.06599","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"pith_short_12","alias_value":"5H7BMODOI5NN","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"pith_short_16","alias_value":"5H7BMODOI5NN2RW5","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"pith_short_8","alias_value":"5H7BMODO","created_at":"2026-07-05T09:45:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2021:5H7BMODOI5NN2RW5CGOVG4ACJ4","target":"record","payload":{"canonical_record":{"source":{"id":"2104.06599","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-04-14T02:56:14Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"278a8260ad95ea1ce912af58cf13afff65c148cfde31ed3d6e26bbcbc3292f43","abstract_canon_sha256":"5a0528f100109f5df7b47d76aa10a4157b0f23ec375c6afee9a3877580158731"},"schema_version":"1.0"},"canonical_sha256":"e9fe16386e475add46dd119d5370024f11ffdb5e0c7fb3c207e540630ebecbdf","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:45:36.354496Z","signature_b64":"ZN5yFCThM2ppjrJQ8MVvkQ/sNJ3QsFkguPAJ8vtoQEB9dVDufIAND28pFxRc5DZrwE/CuLWZzryM9j+ebrIkBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e9fe16386e475add46dd119d5370024f11ffdb5e0c7fb3c207e540630ebecbdf","last_reissued_at":"2026-07-05T09:45:36.353954Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:45:36.353954Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2104.06599","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:45:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZCjBtyC2p44mmaYaJPk/LXRRpOYOGZe1jtUbFTBwxnz0XouLjRQBdYHZxdpCai5OneWX/MsaEOZg5FmO156WDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T00:03:00.425766Z"},"content_sha256":"501498342ead1a767016feda4a8109b93361bd5a677da5577d21e388d0183d42","schema_version":"1.0","event_id":"sha256:501498342ead1a767016feda4a8109b93361bd5a677da5577d21e388d0183d42"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2021:5H7BMODOI5NN2RW5CGOVG4ACJ4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning How to Ask: Querying LMs with Mixtures of Soft Prompts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Guanghui Qin, Jason Eisner","submitted_at":"2021-04-14T02:56:14Z","abstract_excerpt":"Natural-language prompts have recently been used to coax pretrained language models into performing other AI tasks, using a fill-in-the-blank paradigm (Petroni et al., 2019) or a few-shot extrapolation paradigm (Brown et al., 2020). For example, language models retain factual knowledge from their training corpora that can be extracted by asking them to \"fill in the blank\" in a sentential prompt. However, where does this prompt come from? We explore the idea of learning prompts by gradient descent -- either fine-tuning prompts taken from previous work, or starting from random initialization. Ou"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2104.06599","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2104.06599/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:45:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TS1qYq8bINgbshCF4OX/AaO5k0/q7bz2lXFL9+pvD+Rkv7TvSpfUlC4Rz7Rzd2c1/Q7FiQEnHD3Ags9NqmXuBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T00:03:00.426073Z"},"content_sha256":"ee2490b923662d5b1931de0c6f0a30f2929576c71d833e9cff7786bd04041058","schema_version":"1.0","event_id":"sha256:ee2490b923662d5b1931de0c6f0a30f2929576c71d833e9cff7786bd04041058"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5H7BMODOI5NN2RW5CGOVG4ACJ4/bundle.json","state_url":"https://pith.science/pith/5H7BMODOI5NN2RW5CGOVG4ACJ4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5H7BMODOI5NN2RW5CGOVG4ACJ4/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-15T00:03:00Z","links":{"resolver":"https://pith.science/pith/5H7BMODOI5NN2RW5CGOVG4ACJ4","bundle":"https://pith.science/pith/5H7BMODOI5NN2RW5CGOVG4ACJ4/bundle.json","state":"https://pith.science/pith/5H7BMODOI5NN2RW5CGOVG4ACJ4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5H7BMODOI5NN2RW5CGOVG4ACJ4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2021:5H7BMODOI5NN2RW5CGOVG4ACJ4","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"5a0528f100109f5df7b47d76aa10a4157b0f23ec375c6afee9a3877580158731","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-04-14T02:56:14Z","title_canon_sha256":"278a8260ad95ea1ce912af58cf13afff65c148cfde31ed3d6e26bbcbc3292f43"},"schema_version":"1.0","source":{"id":"2104.06599","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2104.06599","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"arxiv_version","alias_value":"2104.06599v1","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2104.06599","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"pith_short_12","alias_value":"5H7BMODOI5NN","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"pith_short_16","alias_value":"5H7BMODOI5NN2RW5","created_at":"2026-07-05T09:45:36Z"},{"alias_kind":"pith_short_8","alias_value":"5H7BMODO","created_at":"2026-07-05T09:45:36Z"}],"graph_snapshots":[{"event_id":"sha256:ee2490b923662d5b1931de0c6f0a30f2929576c71d833e9cff7786bd04041058","target":"graph","created_at":"2026-07-05T09:45:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2104.06599/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Natural-language prompts have recently been used to coax pretrained language models into performing other AI tasks, using a fill-in-the-blank paradigm (Petroni et al., 2019) or a few-shot extrapolation paradigm (Brown et al., 2020). For example, language models retain factual knowledge from their training corpora that can be extracted by asking them to \"fill in the blank\" in a sentential prompt. However, where does this prompt come from? We explore the idea of learning prompts by gradient descent -- either fine-tuning prompts taken from previous work, or starting from random initialization. Ou","authors_text":"Guanghui Qin, Jason Eisner","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-04-14T02:56:14Z","title":"Learning How to Ask: Querying LMs with Mixtures of Soft Prompts"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2104.06599","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:501498342ead1a767016feda4a8109b93361bd5a677da5577d21e388d0183d42","target":"record","created_at":"2026-07-05T09:45:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"5a0528f100109f5df7b47d76aa10a4157b0f23ec375c6afee9a3877580158731","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-04-14T02:56:14Z","title_canon_sha256":"278a8260ad95ea1ce912af58cf13afff65c148cfde31ed3d6e26bbcbc3292f43"},"schema_version":"1.0","source":{"id":"2104.06599","kind":"arxiv","version":1}},"canonical_sha256":"e9fe16386e475add46dd119d5370024f11ffdb5e0c7fb3c207e540630ebecbdf","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e9fe16386e475add46dd119d5370024f11ffdb5e0c7fb3c207e540630ebecbdf","first_computed_at":"2026-07-05T09:45:36.353954Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:45:36.353954Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ZN5yFCThM2ppjrJQ8MVvkQ/sNJ3QsFkguPAJ8vtoQEB9dVDufIAND28pFxRc5DZrwE/CuLWZzryM9j+ebrIkBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T09:45:36.354496Z","signed_message":"canonical_sha256_bytes"},"source_id":"2104.06599","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:501498342ead1a767016feda4a8109b93361bd5a677da5577d21e388d0183d42","sha256:ee2490b923662d5b1931de0c6f0a30f2929576c71d833e9cff7786bd04041058"],"state_sha256":"cce0ccc2a58cea5051f6b92eaf511da0fdeb8f09b02b2abab2974e6dd549fc10"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Mi4qUjdH9dHdTmQQvPoyeTwUD5crUrH1jAsvc5mEdb9zMrb7R5hkhtALeHKaWfcZau9syqTycZzTC3obPfZSAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-15T00:03:00.430105Z","bundle_sha256":"1cb81ab8962998f5472f385162852cef78e2237d09047db2f6ebb079fff32335"}}