{"paper":{"title":"Releasing the Inner Jev: A Simple Recipe for Open Decision Models","arxiv_id":"2610.17205","primary_cat":"cs.AI","submitted_at":"2026-10-12T17:59:37+00:00","version_compiled":"v1"},"notice":"Every quoted sentence is the authors' text from the arXiv listing abstract. Labels and probabilities are the outputs of classifier-1.13.0 on the question bank abstract-p1-20260923, exactly as returned. No sentence in this record was written by a language model.","schema":"pith.compiled-record/v0","status":"compiled","abstract":{"source":"arXiv listing abstract v1","sentences":[{"i":0,"p":1.0,"id":"s0","own":false,"role":"background","text":"LLM-based decision-making systems, also called System One models, have recently gained popularity, most notably through Jev (Almeida, 2026).","hedge":{"p":1.0,"label":"asserted"},"own_p":0.04,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.12,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":1.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":1,"p":0.97,"id":"s1","own":false,"role":"background","text":"Given a typed input state, a question, and a set of options, they return a probability for each option in a single forward pass, without text generation.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.06,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.15,"role_probabilities":{"gap":0.0,"method":0.03,"result":0.0,"background":0.97,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":2,"p":0.61,"id":"s2","own":true,"role":"contribution","text":"We show that existing open instruction-tuned language models can achieve comparable performance \\textit{without any additional training}, by reading out answers from letters (A, B, C, etc.) in the model's output logits at the last token position.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.91,"main_claim_p":0.3,"quantitative":false,"quantitative_p":0.11,"role_probabilities":{"gap":0.0,"method":0.23,"result":0.16,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.61,"interpretation":0.0}},{"i":3,"p":0.66,"id":"s3","own":true,"role":"contribution","text":"Furthermore, we show how a simple supervised fine-tuning recipe can improve their performance further.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.85,"main_claim_p":0.07,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.33,"result":0.01,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.66,"interpretation":0.0}},{"i":4,"p":0.96,"id":"s4","own":true,"role":"method","text":"We evaluate four open instruction-tuned language models: Gemma 4 31B, Gemma 4 12B, Qwen3.6-27B, and Qwen3.5-4B.","hedge":{"p":0.85,"label":"asserted"},"own_p":0.86,"main_claim_p":0.02,"quantitative":true,"quantitative_p":0.69,"role_probabilities":{"gap":0.0,"method":0.96,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.04,"interpretation":0.0}},{"i":5,"p":1.0,"id":"s5","own":true,"role":"method","text":"We fine-tune each using LoRA and a decision head initialized from a slice of the language model head.","hedge":{"p":0.91,"label":"asserted"},"own_p":0.95,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.0,"method":1.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":6,"p":0.68,"id":"s6","own":true,"role":"contribution","text":"Furthermore, we introduce OpenDecisions, a novel 2,352-question validation dataset designed to reflect the task composition of JevBench (JevBench main- tainers, 2026) and which computes a Capability score based on accuracy and calibration metrics.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.95,"main_claim_p":0.04,"quantitative":true,"quantitative_p":0.95,"role_probabilities":{"gap":0.0,"method":0.32,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.68,"interpretation":0.0}},{"i":7,"p":1.0,"id":"s7","own":true,"role":"result","text":"Without fine-tuning, Gemma 4 31B scores 79.6 on OpenDecisions, compared to Jev's 80.1.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.89,"main_claim_p":0.06,"quantitative":true,"quantitative_p":0.99,"role_probabilities":{"gap":0.0,"method":0.0,"result":1.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":8,"p":1.0,"id":"s8","own":true,"role":"result","text":"After undergoing our fine-tuning recipe, it scores 84.4 on OpenDecisions and 81.5 on JevBench v1.7, outperforming Jev by 4.3 and 4.4 Capability points respectively and setting a new state-of-the-art on OpenDecisions.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.97,"main_claim_p":0.2,"quantitative":true,"quantitative_p":0.99,"role_probabilities":{"gap":0.0,"method":0.0,"result":1.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":9,"p":1.0,"id":"s9","own":true,"role":"result","text":"On a suite of public decision datasets, our models achieve accuracy comparable to Jev, with lower calibration error.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.95,"main_claim_p":0.13,"quantitative":true,"quantitative_p":0.87,"role_probabilities":{"gap":0.0,"method":0.0,"result":1.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":10,"p":0.97,"id":"s10","own":true,"role":"interpretation","text":"Our results show that open language models already possess System One capabilities, and with simple fine-tuning they outperform Jev on a wide range of decision tasks.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.95,"main_claim_p":0.93,"quantitative":true,"quantitative_p":0.86,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.03,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.97}},{"i":11,"p":0.99,"id":"s11","own":true,"role":"contribution","text":"We release our models, serving code, and OpenDecisions.","hedge":{"p":0.79,"label":"asserted"},"own_p":0.8,"main_claim_p":0.04,"quantitative":false,"quantitative_p":0.03,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.01,"contribution":0.99,"interpretation":0.0}}]},"compiled":{"main_claim":{"p":1.0,"text":"Our results show that open language models already possess System One capabilities, and with simple fine-tuning they outperform the classifier on a wide range of decision tasks.","location":"abstract s10","hedge_strength":{"p":1.0,"label":"asserted"}},"provenance":{"route":{"via":"TypeSafe API","model":"classifier-1.13.0","endpoint":"https://api.typesafe.ai/v1/systemone","provider":"TypeSafe"},"packing":"one request per paper","requests":1,"input_tokens":9175,"cost_usd_reported":0.0},"compiled_at":"2026-10-14","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.09,"states_limitation":0.12,"human_participants":0.04,"controlled_comparison":0.89,"randomized_assignment":0.01,"states_quantitative_result":0.99,"compares_with_prior_results":0.98,"mentions_code_or_data_availability":0.98},"evidential_basis":{"p":0.98,"value":"observation or experiment","probabilities":{"argument":0.0,"literature":0.0,"proof_or_derivation":0.0,"analysis_of_existing_data":0.0,"observation_or_experiment":0.98,"simulation_or_computation":0.02}},"contribution_type":{"p":0.79,"value":"method or tool","probabilities":{"review":0.0,"method_or_tool":0.79,"empirical_study":0.16,"simulation_study":0.0,"theoretical_result":0.0,"dataset_or_benchmark":0.05,"position_or_commentary":0.0}},"contribution_statement":{"p":0.58,"text":"We show that existing open instruction-tuned language models can achieve comparable performance \\textit{without any additional training}, by reading out answers from letters (A, B, C, etc.) in the model's output logits at the last token position.","location":"abstract s2"}},"coverage":{"tier":"abstract","levels_run":["P abstract"],"not_examined":["full text","figures and tables","references and citing sentences","relations to other papers"],"sources_read":["arXiv listing abstract v1"],"levels_not_run":["S sections","T body sentences","C citations","R cross-paper relations"],"sentences_examined":12,"sentences_in_abstract":12},"measured_accuracy":{"url":"https://pith.science/accuracy","measured_on":"2026-10-02","method":"two independent model families labelled a fresh random sample blind; the figure is how often the published label matches their consensus","sentence_role_agreement":0.92,"main_claim_agreement":0.908,"typed_citation_agreement":0.933,"typed_citation_agreement_95_interval":[0.917,0.945],"review_comments_accurate":0.758,"review_comments_accurate_95_interval":[0.684,0.82],"review_caught_admitted_error_found_or_related":0.448,"review_comments_judged":{"judged_by":["gpt-5.6-luna","grok-4.7"],"judged_against":"Pith's own text copy of each paper: text extracted from the PDF, the first 90,000 characters","measured_on":"2026-10-04","accurate":113,"decided":149,"sampled_reviews":200,"split_by_lane":false,"by_lane":"https://pith.science/accuracy#by-lane"}}}