{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/softmax-probabilities-mostly-predict-large","title":"Probabilities of Chat LLMs Are Miscalibrated but Still Predict Correctness on Multiple-Choice Q&A","arxiv_id":"2402.13213","date":"2024-02-20","proceeding":null,"authors":["Benjamin Plaut","Nguyen X. Khanh","Tu Trinh"],"abstract":"We study 14 large language models (LLMs) fine-tuned for chat and find that their maximum softmax probabilities (MSPs) are consistently miscalibrated on multiple-choice Q&A. However, those MSPs might still encode useful uncertainty information. Specifically, we hypothesized that wrong answers would be associated with smaller MSPs compared to correct answers. Via rigororous statistical testing, we show that this hypothesis holds for models which perform well on the underlying Q&A task. We also find a strong direction correlation between Q&A accuracy and MSP correctness prediction, while finding no correlation between Q&A accuracy and calibration error. This suggests that within the current fine-tuning paradigm, we can expect correctness prediction but not calibration to improve as LLM capabilities progress. To demonstrate the utility of correctness prediction, we show that when models have the option to abstain, performance can be improved by selectively abstaining based on the MSP of the initial model response, using only a small amount of labeled data to choose the MSP threshold.","url_abs":"https://arxiv.org/abs/2402.13213v2","url_pdf":"https://arxiv.org/pdf/2402.13213v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"softmax-probabilities-mostly-predict-large","repo_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"softmax-probabilities-mostly-predict-large","repo_url":"https://github.com/bplaut/llm-calibration-and-correctness-prediction","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"large-language-model","task_name":"Large Language Model"},{"task_slug":"multiple-choice","task_name":"Multiple-choice"},{"task_slug":"prediction","task_name":"Prediction"}],"methods":[{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2402.13213","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2402.13213"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/bplaut/llm-calibration-and-correctness-prediction","reach":null}],"summary":{"ran":9,"ran_fixture":1,"unverified":4},"by_repo_kind":{"official":{"samples":14,"ran":10,"repositories":2}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"74f56cba2c962f66","entry":"compute_statistics","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"show_msp_distributions.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/show_msp_distributions.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"74f56cba2c962f66"}},{"code_sha256_prefix":"ed723c89b7e88189","entry":"parse_data","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"utils.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ed723c89b7e88189"}},{"code_sha256_prefix":"54531f830631efef","entry":"parse_file_name","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"utils.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"54531f830631efef"}},{"code_sha256_prefix":"ddaa2dc3f0678532","entry":"parse_filename","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"show_msp_distributions.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/show_msp_distributions.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ddaa2dc3f0678532"}},{"code_sha256_prefix":"0881fdbaf4528342","entry":"parse_group_name","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"utils.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0881fdbaf4528342"}},{"code_sha256_prefix":"44d2ee7d0ca61946","entry":"plot_style_for_group","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"plotting_functions.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/plotting_functions.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"44d2ee7d0ca61946"}},{"code_sha256_prefix":"0c29fde88ebbe1cd","entry":"process_file","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"show_msp_distributions.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/show_msp_distributions.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0c29fde88ebbe1cd"}},{"code_sha256_prefix":"68f6a3ddce1f5422","entry":"t_to_str","repo":"bplaut/llm-calibration-and-correctness-prediction","repo_kind":"official","path":"generate_text.py","file_url":"https://github.com/bplaut/llm-calibration-and-correctness-prediction/blob/HEAD/generate_text.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"68f6a3ddce1f5422"}},{"code_sha256_prefix":"3b63c9c481b3cc15","entry":"test_large_sample","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"statistical_tests.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/statistical_tests.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3b63c9c481b3cc15"}},{"code_sha256_prefix":"95b959aa05271726","entry":"test_normality","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"statistical_tests.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/statistical_tests.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"95b959aa05271726"}},{"code_sha256_prefix":"4d4f6a798379d5f5","entry":"acc_scatter_plots","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"plotting_functions.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/plotting_functions.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4d4f6a798379d5f5"}},{"code_sha256_prefix":"4e2deaf4a017eea3","entry":"compute_auroc","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"plotting_functions.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/plotting_functions.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4e2deaf4a017eea3"}},{"code_sha256_prefix":"8d80acdead51a530","entry":"plots_for_group","repo":"bplaut/llm-calibration-and-correctness-prediction","repo_kind":"official","path":"analysis_main.py","file_url":"https://github.com/bplaut/llm-calibration-and-correctness-prediction/blob/HEAD/analysis_main.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8d80acdead51a530"}},{"code_sha256_prefix":"85a8cea633d8182d","entry":"test_equal_variance","repo":"bplaut/softmax-probs-predict-llm-correctness","repo_kind":"official","path":"statistical_tests.py","file_url":"https://github.com/bplaut/softmax-probs-predict-llm-correctness/blob/HEAD/statistical_tests.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"85a8cea633d8182d"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}