{"_canonicalization":{"envelope_id":"axm_ + sha256(envelope minus {signature, axiom_id, anchors})","envelope_signature":"ed25519(envelope minus {signature, axiom_id})","json":"sort_keys=True, separators=(',',':'), ensure_ascii=False, allow_nan=False, utf-8","leaf_hash":"sha256(0x00 || canonical_json(envelope_full))","seal_signature":"ed25519(seal minus {signature, sig_algorithm})"},"axiom_id":"axm_26ec520df81abde8f252767b31685d15b76120ee9845447cb93d0b1896f06247","bitcoin_anchor":{"bitcoin_attestations":[],"calendar_attestations":[],"ots_url":"","stamped_at":"","status":"pending_next_stamp"},"envelope":{"anchors":[{"chain":"crovia.axiom_graph","height":0,"merkle_proof":"spider_vendor_press_v1","root_at_anchor":"spider_vendor_press_v1"}],"axiom_id":"axm_26ec520df81abde8f252767b31685d15b76120ee9845447cb93d0b1896f06247","axiom_type":"AX.OBS","body":{"axiom_subtype":"news.vendor_press.v1","category":"news","fingerprint":"bd712394075fc3005b0257e8e8385fafe2d6213556902610e1d786b2d73bbcf3","published":"Wed, 24 Jun 2026 00:00:00 -0400","receipt_hash":"bd712394075fc3005b0257e8e8385fafe2d6213556902610e1d786b2d73bbcf3","schema":"spider.news.vendor_press.v1","spider":"vendor_press","spider_record":{"axiom_subtype":"news.vendor_press.v1","category":"news","decision_hint":"POSITIVE","envelope_target":"AX.OBS","fingerprint":"bd712394075fc3005b0257e8e8385fafe2d6213556902610e1d786b2d73bbcf3","observed_at":"2026-06-24T04:43:17.877668Z","parent_run_hash":"ca17d06d44ba7db934e6f913874699efc608b8f87453f1ac67f52060e620b57c","published":"Wed, 24 Jun 2026 00:00:00 -0400","runtime_version":"0.1.0","schema":"spider.news.vendor_press.v1","source_status":200,"source_url":"https://export.arxiv.org/rss/cs.AI","spider":"vendor_press","summary_excerpt":"arXiv:2606.24267v1 Announce Type: cross \nAbstract: While in-context learning is generally shown to be effective in Large Language Models (LLMs), bad contexts can cause performance degradation and mode collapse, a phenomenon we call \"pigeonholing.\" **Unintentionally bad** contexts can happen without malicious jailbreaking intents: For example, a user asks the model to justify an incorrect math theorem or fails to correct the model's buggy code. Specifically, we investigate ``pigeonholing\" in two scenarios: (1) when the user suggests a solution, and (2) when the conversation context includes the assistant's previous (incorrect) responses. Our experiments across 10 verifiable and open-ended tasks with 10 different models show that pigeonholing manifests in several ways: (1) repeating the incorrect answers from context (leading to 38-40% performance drop), (2) converging on a narrow set of answers in coding and text generation without exploring alternatives, and (3) flipping stance on cont","title":"Pigeonholing: Bad prompts hurt models to collapse and make mistakes","url":"https://arxiv.org/abs/2606.24267","vendor":"arxiv_cs_ai"},"summary":"arXiv:2606.24267v1 Announce Type: cross \nAbstract: While in-context learning is generally shown to be effective in Large Language Models (LLMs), bad contexts can cause performance degradation and mode collapse, a phenomenon we call \"pigeonholing.\" **Unintentionally bad** contexts can happen without malicious jailbreaking intents: For example, a user asks the model to justify an incorrect math theorem or fails to correct the model's buggy code. Specifically, we investigate ``pigeonholing\" in two scenarios: (1) when the user suggests a solution, and (2) when the conversation context includes the assistant's previous (incorrect) responses. Our experiments across 10 verifiable and open-ended tasks with 10 different models show that pigeonholing manifests in several ways: (1) repeating the incorrect answers from context (leading to 38-40% performance drop), (2) converging on a narrow set of answers in coding and text generation without exploring alternatives, and (3) flipping stance on cont","title":"Pigeonholing: Bad prompts hurt models to collapse and make mistakes","vendor":"arxiv_cs_ai"},"confidence":{"method":"deterministic"},"decision":"POSITIVE","issued_at":"2026-06-24T04:43:17Z","notes":"Spider vendor_press (news) news.vendor_press.v1","object":{"captured_by":"crovia.spider.vendor_press","primary_source_url":"https://arxiv.org/abs/2606.24267"},"predecessors":[],"schema":"crovia.axiom.v1","signature":"ed25519:e55eefb46a6615257ee1c346867f4281618f83f9f6e46d6e8666b8352b20f4c7defe750790daeb5358df1d2b060754246daf19874b0ea6435b8a564964d1aa0a","signer":"crovia.substrate","subject":{"observed_at":"2026-06-24T04:43:17Z","source_collector":"spider:vendor_press","target_id":"https://arxiv.org/abs/2606.24267"},"tsa":{"authority":"crovia.substrate.bootstrap","rfc3161_token":"{\"kind\":\"crovia.bootstrap.tsa\",\"source_jsonl\":\"/opt/crovia/spider/data/news/vendor_press_v1.jsonl\",\"source_seal_merkle_root\":\"spider_vendor_press_v1\",\"upgrade_path\":\"Sessione H \\u2014 OpenTimestamps weekly anchor\"}"},"zk_mode":"clear","zk_proof":null},"ledger":{"leaf_hash":"8c342bc9e20f23a399061b7b749cb16bab816131043ea416fa771c4fe6fe2ce1","leaf_index":244566,"ledger_path":"/opt/crovia/substrate/axiom_ledger.jsonl"},"merkle_proof":{"hash_alg":"sha256","leaf_prefix":"0x00","node_prefix":"0x01","odd_leaf_rule":"duplicate_last","path":[{"sibling":"a0dba18bef32df72735c34ff024f061c76408aca14bcf50266e76e4dba80d438","side":"right"},{"sibling":"d75cab56ebafadd8389a8c68805829ecdd569c634fe4ef398c3b6b2261e29013","side":"left"},{"sibling":"4f12bc11d384a5978a2399933ac36e2cf8060b7aad78d70631c475c6260d23a4","side":"left"},{"sibling":"1abe02f005cd0d8b0c3cf66807721dec49af0e5e72f31090f215c3e804d8b95b","side":"right"},{"sibling":"92703a887e7f1bf8322f8d44ac0cb59c1842671e6697413180532d2dd09d35d1","side":"left"},{"sibling":"b067534616bb7d7c6554b89a9447081f53dff0fccf744257ad56cf669f711947","side":"right"},{"sibling":"2fb6532259e6852414e5b4ef4dd8ffec5a8a48bee3324303be4c62aabee72329","side":"left"},{"sibling":"d4fcf6bcc7f1fe78997b28df74b23bba21f5b2585d6fe8efb802f0d067291d5c","side":"right"},{"sibling":"0fa23771b702ff726ed1fc5a44f9b416b2a7861c2f957fcac2d95392276d4784","side":"left"},{"sibling":"bc74ebb08462da8a50fc65ea75f8a8a3418d10ebd471d830f1c67f33dd54dfd1","side":"left"},{"sibling":"6dafd355e5d54c60e61c6c02d3842984e234b1f5bca1623fcdd3def7b8931973","side":"right"},{"sibling":"3107b9d4dbf9456a39f99de694a4dd4da2c0600f9f8855f125161335fe8810af","side":"left"},{"sibling":"86118ab4500c3055a2af70062751a960423c464405b18ca1c37411bf0ce3f52e","side":"left"},{"sibling":"3a42039065acac6d3e4088ec61d9c116ecf7a26c7b7163d23da8fd0b3362e038","side":"left"},{"sibling":"c044f2bd864a0e8e8af5a7f6e3124def7fc4b4511b2b166ea8f9de321e8d385e","side":"right"},{"sibling":"a04392fb9f2a3a620840e3b3fecd93d12e6d224e481c162389d8ae64e7194599","side":"left"},{"sibling":"c300cf0154c136afc09b1702a0be98f4ba5b6dc5cf57e8cc714ec1eaf4196eff","side":"left"},{"sibling":"d841ad93efda0869e5eb97678f348f03f5caab4353e05ff4bf18f47fb945b822","side":"left"}]},"schema":"crovia.axiom_proof.v1","seal":{"first_collector_run_id":"","first_receipt_hash":"","jsonl_path":"/opt/crovia/substrate/axiom_ledger.jsonl","key_id":"430895f101d38164","last_collector_run_id":"","last_receipt_hash":"","leaf_count":244827,"merkle_root":"274e133c6dfa2781a9cfb85337d01cc6b72688ce5e810149f3183e400ffab136","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","run_id":"hourly_json_retrofit_20260624T053701Z","schema":"crovia.seal.v1","seal_family_version":"crovia-seal-family/1","seal_kind":"substrate_batch","sealed_at":"2026-06-24T05:37:55Z","sig_algorithm":"ed25519","signature":"22ca3cee4de2447b3d281e30e09fe566461996bb7be4d4465f08a3f4cf59cea58f22683a4aa10ef4d5d17a19b03f4212392bfd26f2b51f289f0cdf1042a03800","signer_version":"1.1.0"},"trust_root":{"key_id":"430895f101d38164","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","signature_algorithm":"ed25519","url":"/registry/canon/TRUST_ROOT.md"},"verifier":{"spec":"/registry/canon/AXIOM_RECEIPT_v1.md","url":"/v/axm_26ec520df81abde8f252767b31685d15b76120ee9845447cb93d0b1896f06247"}}