{"work":{"id":"9bce40c8-cfd7-4983-80e0-c3bd4402322a","openalex_id":null,"doi":null,"arxiv_id":"2302.04761","raw_key":null,"title":"Toolformer: Language Models Can Teach Themselves to Use Tools","authors":null,"authors_text":"Timo Schick, Jane Dwivedi-Yu, Roberto Dess\\`i, Roberta Raileanu, Maria Lomeli, Luke Zettlemoyer","year":2023,"venue":"cs.CL","abstract":"Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with basic functionality, such as arithmetic or factual lookup, where much simpler and smaller models excel. In this paper, we show that LMs can teach themselves to use external tools via simple APIs and achieve the best of both worlds. We introduce Toolformer, a model trained to decide which APIs to call, when to call them, what arguments to pass, and how to best incorporate the results into future token prediction. This is done in a self-supervised way, requiring nothing more than a handful of demonstrations for each API. We incorporate a range of tools, including a calculator, a Q\\&A system, two different search engines, a translation system, and a calendar. Toolformer achieves substantially improved zero-shot performance across a variety of downstream tasks, often competitive with much larger models, without sacrificing its core language modeling abilities.","external_url":"https://arxiv.org/abs/2302.04761","cited_by_count":null,"metadata_source":"pith","metadata_fetched_at":"2026-06-28T22:22:43.404876+00:00","pith_arxiv_id":"2302.04761","created_at":"2026-05-08T22:14:19.973855+00:00","updated_at":"2026-06-28T22:22:43.404876+00:00","title_quality_ok":true,"display_title":"Toolformer: Language Models Can Teach Themselves to Use Tools","render_title":"Toolformer: Language Models Can Teach Themselves to Use Tools"},"hub":{"state":{"work_id":"9bce40c8-cfd7-4983-80e0-c3bd4402322a","tier":"super_hub","tier_reason":"100+ Pith inbound or 10,000+ external citations","pith_inbound_count":145,"external_cited_by_count":null,"distinct_field_count":17,"first_pith_cited_at":"2022-11-22T21:06:00+00:00","last_pith_cited_at":"2026-06-21T13:34:18+00:00","author_build_status":"needed","summary_status":"needed","contexts_status":"needed","graph_status":"needed","ask_index_status":"needed","reader_status":"not_needed","recognition_status":"not_needed","updated_at":"2026-06-29T11:58:42.027984+00:00","tier_text":"super_hub"},"tier":"super_hub","role_counts":[{"context_role":"background","n":44},{"context_role":"baseline","n":1},{"context_role":"dataset","n":1},{"context_role":"method","n":1}],"polarity_counts":[{"context_polarity":"background","n":41},{"context_polarity":"support","n":2},{"context_polarity":"unclear","n":2},{"context_polarity":"baseline","n":1},{"context_polarity":"use_method","n":1}],"runs":{"ask_index":{"job_type":"ask_index","status":"succeeded","result":{"title":"Toolformer: Language Models Can Teach Themselves to Use Tools","claims":[{"claim_text":"Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with basic functionality, such as arithmetic or factual lookup, where much simpler and smaller models excel. In this paper, we show that LMs can teach themselves to use external tools via simple APIs and achieve the best of both worlds. We introduce Toolformer, a model trained to decide which APIs to call, when to call them, what arguments to pass, and how to best incorporate the results into future token prediction. Thi","claim_type":"abstract","evidence_strength":"source_metadata"},{"claim_text":"This work is expected to serve as a critical wake-up call for the security of the rapidly evolving AI-agentic infrastructure. I. INTRODUCTION The rapid evolution of Large Language Models (LLMs) has catalyzed the transition from passive chatbots to autonomous LLM Agents [1], [2]. These agents, represented by Claude Code [3], OpenClaw [4], and various frameworks [5], [6], are characterized by their ability to perform complex reasoning and execute actions via external \"skills\" or tools [7], [6]. By","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"arXiv e-printsarXiv:2201.11903 (2022), doi:10.48550/arXiv.2201.11903. [13] S. Yao,et al., ReAct: Synergizing Reasoning and Acting in Language Models.arXiv e-printsarXiv:2210.03629 (2022), doi:10.48550/arXiv.2210.03629. [14] T. Schick,et al., Toolformer: Language Models Can Teach Themselves to Use Tools.arXiv e-printsarXiv:2302.04761 (2023), doi:10.48550/arXiv.2302.04761. [15] F. Villaescusa-Navarro,et al., The Denario project: Deep knowledge AI agents for scientific discovery.ArXivabs/2510.26887","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"\"Evaluation\" indicates whether the model has been evaluated with corresponding abilities in their original paper: ICL denotes in-context learning and CoT denotes chain-of-thought. \"*\" denotes the largest publicly available version. Adaptation EvaluationModel Release Time Size (B) Base Model IT RLHF Pre-train Data Scale Latest Data Timestamp Hardware (GPUs / TPUs) Training Time ICL CoT T5 [82] Oct-2019 11 - - - 1T tokens Apr-2019 1024 TPU v3 -✓- mT5 [83] Oct-2020 13 - - - 1T tokens - - -✓- PanGu-","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"A stateless LLM responds to a single input and produces a single output, creating a bounded, session-isolated attack surface amenable to traditional input/output filtering. An AI agent, by contrast, maintains persistent state across sessions, formulates multi-step plans, invokes privileged external tools (web browsers, code interpreters, file systems, APIs) [1]- [4], and increasingly coordinates with other autonomous agents to accomplish long-horizon tasks [5]-[9]. This shift is driven partly by","claim_type":"background","confidence":0.85,"evidence_strength":"citation_context"},{"claim_text":"in a self-reinforcing cycle. Second, to overcome the difficulties ofscalable multi-turn RL, we design a training framework that stabilizes optimization in long-horizon settings. This includes asynchronous rollouts with stateful environments to preserve context, streaming updates to avoid bottlenecks from long-tail trajectories, and enhanced proximal policy optimization [55] with reward shaping, adaptive advantage estimation, and value pretraining. Third, to move beyond the limitations of pure GU","claim_type":"background","confidence":0.8,"evidence_strength":"citation_context"},{"claim_text":"anda t is the resulting action. Recent systems such as Open- Claw and Manus exemplify this shift from passive response generation to action-oriented execution [7], [8]. What separates modern agents from standalone LLMs is that they canact: they query external systems, invoke tools, write and execute code, and coordinate with other agents [5], [6], [21]. That action capability is the foundation of the capability stack studied in this survey. B. Agent Knowledge and the Procedural Gap Agent behavio","claim_type":"background","confidence":0.8,"evidence_strength":"citation_context"}],"why_cited":"Pith tracks Toolformer: Language Models Can Teach Themselves to Use Tools because it crossed a citation-hub threshold. Current citing contexts most often use it as background evidence (8 contexts).","role_counts":[{"n":8,"context_role":"background"}]},"error":null,"updated_at":"2026-05-15T11:47:49.694398+00:00"},"author_expand":{"job_type":"author_expand","status":"succeeded","result":{"authors_linked":[{"id":"732acc80-1625-494f-9f40-0563e525decb","orcid":null,"display_name":"Timo Schick"},{"id":"f9a9a5c9-d960-4ac5-81e1-8cb2f89f8f35","orcid":null,"display_name":"Jane Dwivedi-Yu"},{"id":"99ca19d6-b858-45a4-a2e3-602e89e3c4c2","orcid":null,"display_name":"Roberto Dess\\`i"},{"id":"a2e4441f-90f0-4b2e-8417-97fa35eb28d8","orcid":null,"display_name":"Roberta Raileanu"},{"id":"f778ed73-e866-4f46-a108-bc66cc2ce2f2","orcid":null,"display_name":"Maria Lomeli"},{"id":"383382cc-d254-4d3b-92b6-03e2ff131701","orcid":null,"display_name":"Luke Zettlemoyer"}]},"error":null,"updated_at":"2026-05-15T11:47:45.092624+00:00"},"context_extract":{"job_type":"context_extract","status":"succeeded","result":{"enqueued_papers":25},"error":null,"updated_at":"2026-05-14T06:47:36.470492+00:00"},"graph_features":{"job_type":"graph_features","status":"succeeded","result":{"co_cited":[{"title":"ReAct: Synergizing Reasoning and Acting in Language Models","work_id":"407a2351-25f1-497d-b611-f77d0292a8e6","shared_citers":46},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":19},{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":18},{"title":"Voyager: An Open-Ended Embodied Agent with Large Language Models","work_id":"ffe0d207-86cf-4742-a100-e988ac8b9676","shared_citers":17},{"title":"Reflexion: Language Agents with Verbal Reinforcement Learning","work_id":"778f739e-5f55-4961-8a2a-e4736a2757f4","shared_citers":16},{"title":"AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation","work_id":"92b7eb9c-c3d8-4518-a376-06fa15dd895b","shared_citers":15},{"title":"WebGPT: Browser-assisted question-answering with human feedback","work_id":"e25ef3e1-4848-4cb9-bf28-67a420591165","shared_citers":14},{"title":"WebArena: A Realistic Web Environment for Building Autonomous Agents","work_id":"7058ffd2-a339-4102-89eb-248eeb074652","shared_citers":12},{"title":"AgentBench: Evaluating LLMs as Agents","work_id":"a37549b4-4c94-412d-acc4-4efeb08509be","shared_citers":11},{"title":"Chain-of-Thought Prompting Elicits Reasoning in Large Language Models","work_id":"d1cf6693-a082-403c-ada9-dac7b96341f9","shared_citers":11},{"title":"Gorilla: Large Language Model Connected with Massive APIs","work_id":"126a464a-4a73-495f-b669-de1e44aa8f09","shared_citers":10},{"title":"PaLM: Scaling Language Modeling with Pathways","work_id":"a94f3ef7-2c49-4445-93fe-6ec16aafd966","shared_citers":10},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":10},{"title":"Self-Refine: Iterative Refinement with Self-Feedback","work_id":"59181e7f-e58e-45d3-8146-4477a9f53d5a","shared_citers":10},{"title":"ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs","work_id":"3c555b48-a4d9-42dd-9fdd-0f6018fbe9cb","shared_citers":10},{"title":"A survey on large language model based autonomous agents","work_id":"47f7e8a3-3732-4530-b412-d9c984ce99ed","shared_citers":9},{"title":"Generative Agents: Interactive Simulacra of Human Behavior","work_id":"01f7ddaa-284a-441a-be87-921aad4dc54b","shared_citers":9},{"title":"HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face","work_id":"f20ed1da-2676-4598-a11b-54549718735b","shared_citers":9},{"title":"MRKL systems: A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning","work_id":"e393ecfc-aa97-4014-bb7a-90ebd9f00535","shared_citers":9},{"title":"Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks","work_id":"27eaec54-c105-4969-8188-da5f0fca3688","shared_citers":9},{"title":"SWE-bench: Can Language Models Resolve Real-World GitHub Issues?","work_id":"d0effe15-a689-441a-8e3f-ea35f1c4e4b1","shared_citers":9},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":9},{"title":"Tree of Thoughts: Deliberate Problem Solving with Large Language Models","work_id":"07adb06e-4ed5-4ec5-a7ae-ff288fd214fb","shared_citers":9},{"title":"Llama 2: Open Foundation and Fine-Tuned Chat Models","work_id":"68a5177f-d644-44c1-bd4f-4e5278c22f5d","shared_citers":8}],"time_series":[{"n":1,"year":2022},{"n":13,"year":2023},{"n":2,"year":2024},{"n":1,"year":2025},{"n":70,"year":2026}],"dependency_candidates":[]},"error":null,"updated_at":"2026-05-14T06:47:40.819777+00:00"},"identity_refresh":{"job_type":"identity_refresh","status":"succeeded","result":{"items":[{"title":"Qwen3 Technical Report","outcome":"unchanged","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","resolver":"local_arxiv","confidence":0.98,"old_work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e"}],"counts":{"fixed":0,"merged":0,"unchanged":1,"quarantined":0,"needs_external_resolution":0},"errors":[],"attempted":1},"error":null,"updated_at":"2026-05-14T06:47:45.026641+00:00"},"role_polarity":{"job_type":"role_polarity","status":"succeeded","result":{"title":"Toolformer: Language Models Can Teach Themselves to Use Tools","claims":[{"claim_text":"Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with basic functionality, such as arithmetic or factual lookup, where much simpler and smaller models excel. In this paper, we show that LMs can teach themselves to use external tools via simple APIs and achieve the best of both worlds. We introduce Toolformer, a model trained to decide which APIs to call, when to call them, what arguments to pass, and how to best incorporate the results into future token prediction. Thi","claim_type":"abstract","evidence_strength":"source_metadata"},{"claim_text":"This work is expected to serve as a critical wake-up call for the security of the rapidly evolving AI-agentic infrastructure. I. INTRODUCTION The rapid evolution of Large Language Models (LLMs) has catalyzed the transition from passive chatbots to autonomous LLM Agents [1], [2]. These agents, represented by Claude Code [3], OpenClaw [4], and various frameworks [5], [6], are characterized by their ability to perform complex reasoning and execute actions via external \"skills\" or tools [7], [6]. By","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"arXiv e-printsarXiv:2201.11903 (2022), doi:10.48550/arXiv.2201.11903. [13] S. Yao,et al., ReAct: Synergizing Reasoning and Acting in Language Models.arXiv e-printsarXiv:2210.03629 (2022), doi:10.48550/arXiv.2210.03629. [14] T. Schick,et al., Toolformer: Language Models Can Teach Themselves to Use Tools.arXiv e-printsarXiv:2302.04761 (2023), doi:10.48550/arXiv.2302.04761. [15] F. Villaescusa-Navarro,et al., The Denario project: Deep knowledge AI agents for scientific discovery.ArXivabs/2510.26887","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"\"Evaluation\" indicates whether the model has been evaluated with corresponding abilities in their original paper: ICL denotes in-context learning and CoT denotes chain-of-thought. \"*\" denotes the largest publicly available version. Adaptation EvaluationModel Release Time Size (B) Base Model IT RLHF Pre-train Data Scale Latest Data Timestamp Hardware (GPUs / TPUs) Training Time ICL CoT T5 [82] Oct-2019 11 - - - 1T tokens Apr-2019 1024 TPU v3 -✓- mT5 [83] Oct-2020 13 - - - 1T tokens - - -✓- PanGu-","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"A stateless LLM responds to a single input and produces a single output, creating a bounded, session-isolated attack surface amenable to traditional input/output filtering. An AI agent, by contrast, maintains persistent state across sessions, formulates multi-step plans, invokes privileged external tools (web browsers, code interpreters, file systems, APIs) [1]- [4], and increasingly coordinates with other autonomous agents to accomplish long-horizon tasks [5]-[9]. This shift is driven partly by","claim_type":"background","confidence":0.85,"evidence_strength":"citation_context"},{"claim_text":"in a self-reinforcing cycle. Second, to overcome the difficulties ofscalable multi-turn RL, we design a training framework that stabilizes optimization in long-horizon settings. This includes asynchronous rollouts with stateful environments to preserve context, streaming updates to avoid bottlenecks from long-tail trajectories, and enhanced proximal policy optimization [55] with reward shaping, adaptive advantage estimation, and value pretraining. Third, to move beyond the limitations of pure GU","claim_type":"background","confidence":0.8,"evidence_strength":"citation_context"},{"claim_text":"anda t is the resulting action. Recent systems such as Open- Claw and Manus exemplify this shift from passive response generation to action-oriented execution [7], [8]. What separates modern agents from standalone LLMs is that they canact: they query external systems, invoke tools, write and execute code, and coordinate with other agents [5], [6], [21]. That action capability is the foundation of the capability stack studied in this survey. B. Agent Knowledge and the Procedural Gap Agent behavio","claim_type":"background","confidence":0.8,"evidence_strength":"citation_context"}],"why_cited":"Pith tracks Toolformer: Language Models Can Teach Themselves to Use Tools because it crossed a citation-hub threshold. Current citing contexts most often use it as background evidence (8 contexts).","role_counts":[{"n":8,"context_role":"background"}]},"error":null,"updated_at":"2026-05-15T11:47:49.691671+00:00"},"summary_claims":{"job_type":"summary_claims","status":"succeeded","result":{"title":"Toolformer: Language Models Can Teach Themselves to Use Tools","claims":[{"claim_text":"Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with basic functionality, such as arithmetic or factual lookup, where much simpler and smaller models excel. In this paper, we show that LMs can teach themselves to use external tools via simple APIs and achieve the best of both worlds. We introduce Toolformer, a model trained to decide which APIs to call, when to call them, what arguments to pass, and how to best incorporate the results into future token prediction. Thi","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks Toolformer: Language Models Can Teach Themselves to Use Tools because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-14T06:47:24.212523+00:00"}},"summary":{"title":"Toolformer: Language Models Can Teach Themselves to Use Tools","claims":[{"claim_text":"Language models (LMs) exhibit remarkable abilities to solve new tasks from just a few examples or textual instructions, especially at scale. They also, paradoxically, struggle with basic functionality, such as arithmetic or factual lookup, where much simpler and smaller models excel. In this paper, we show that LMs can teach themselves to use external tools via simple APIs and achieve the best of both worlds. We introduce Toolformer, a model trained to decide which APIs to call, when to call them, what arguments to pass, and how to best incorporate the results into future token prediction. Thi","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks Toolformer: Language Models Can Teach Themselves to Use Tools because it crossed a citation-hub threshold.","role_counts":[]},"graph":{"co_cited":[{"title":"ReAct: Synergizing Reasoning and Acting in Language Models","work_id":"407a2351-25f1-497d-b611-f77d0292a8e6","shared_citers":46},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":19},{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":18},{"title":"Voyager: An Open-Ended Embodied Agent with Large Language Models","work_id":"ffe0d207-86cf-4742-a100-e988ac8b9676","shared_citers":17},{"title":"Reflexion: Language Agents with Verbal Reinforcement Learning","work_id":"778f739e-5f55-4961-8a2a-e4736a2757f4","shared_citers":16},{"title":"AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation","work_id":"92b7eb9c-c3d8-4518-a376-06fa15dd895b","shared_citers":15},{"title":"WebGPT: Browser-assisted question-answering with human feedback","work_id":"e25ef3e1-4848-4cb9-bf28-67a420591165","shared_citers":14},{"title":"WebArena: A Realistic Web Environment for Building Autonomous Agents","work_id":"7058ffd2-a339-4102-89eb-248eeb074652","shared_citers":12},{"title":"AgentBench: Evaluating LLMs as Agents","work_id":"a37549b4-4c94-412d-acc4-4efeb08509be","shared_citers":11},{"title":"Chain-of-Thought Prompting Elicits Reasoning in Large Language Models","work_id":"d1cf6693-a082-403c-ada9-dac7b96341f9","shared_citers":11},{"title":"Gorilla: Large Language Model Connected with Massive APIs","work_id":"126a464a-4a73-495f-b669-de1e44aa8f09","shared_citers":10},{"title":"PaLM: Scaling Language Modeling with Pathways","work_id":"a94f3ef7-2c49-4445-93fe-6ec16aafd966","shared_citers":10},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":10},{"title":"Self-Refine: Iterative Refinement with Self-Feedback","work_id":"59181e7f-e58e-45d3-8146-4477a9f53d5a","shared_citers":10},{"title":"ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs","work_id":"3c555b48-a4d9-42dd-9fdd-0f6018fbe9cb","shared_citers":10},{"title":"A survey on large language model based autonomous agents","work_id":"47f7e8a3-3732-4530-b412-d9c984ce99ed","shared_citers":9},{"title":"Generative Agents: Interactive Simulacra of Human Behavior","work_id":"01f7ddaa-284a-441a-be87-921aad4dc54b","shared_citers":9},{"title":"HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face","work_id":"f20ed1da-2676-4598-a11b-54549718735b","shared_citers":9},{"title":"MRKL systems: A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning","work_id":"e393ecfc-aa97-4014-bb7a-90ebd9f00535","shared_citers":9},{"title":"Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks","work_id":"27eaec54-c105-4969-8188-da5f0fca3688","shared_citers":9},{"title":"SWE-bench: Can Language Models Resolve Real-World GitHub Issues?","work_id":"d0effe15-a689-441a-8e3f-ea35f1c4e4b1","shared_citers":9},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":9},{"title":"Tree of Thoughts: Deliberate Problem Solving with Large Language Models","work_id":"07adb06e-4ed5-4ec5-a7ae-ff288fd214fb","shared_citers":9},{"title":"Llama 2: Open Foundation and Fine-Tuned Chat Models","work_id":"68a5177f-d644-44c1-bd4f-4e5278c22f5d","shared_citers":8}],"time_series":[{"n":1,"year":2022},{"n":13,"year":2023},{"n":2,"year":2024},{"n":1,"year":2025},{"n":70,"year":2026}],"dependency_candidates":[]},"authors":[{"id":"f9a9a5c9-d960-4ac5-81e1-8cb2f89f8f35","orcid":null,"display_name":"Jane Dwivedi-Yu","source":"manual","import_confidence":0.72},{"id":"383382cc-d254-4d3b-92b6-03e2ff131701","orcid":null,"display_name":"Luke Zettlemoyer","source":"manual","import_confidence":0.72},{"id":"f778ed73-e866-4f46-a108-bc66cc2ce2f2","orcid":null,"display_name":"Maria Lomeli","source":"manual","import_confidence":0.72},{"id":"a2e4441f-90f0-4b2e-8417-97fa35eb28d8","orcid":null,"display_name":"Roberta Raileanu","source":"manual","import_confidence":0.72},{"id":"99ca19d6-b858-45a4-a2e3-602e89e3c4c2","orcid":null,"display_name":"Roberto Dess\\`i","source":"manual","import_confidence":0.72},{"id":"732acc80-1625-494f-9f40-0563e525decb","orcid":null,"display_name":"Timo Schick","source":"manual","import_confidence":0.72}]}}