Light Dark GPT-4 / ChatGPT4 / GPT-4 Code Interpreter / GPT-4 Technical Report OpenAI · 2023-03 · text, image · generative · anchor
Closed-weight multimodal language model served through an API, used across the corpus as the strongest general-purpose baseline, including its vision-enabled release.
Variants GPT-4 0613 , GPT-4V / GPT-4 vision , GPT-4 Turbo , GPT-4-turbo-20240409 , GPT-4-0314 , GPT-4-1106 , GPT-4-turbo-0125 , GPT-4-turbo-0409 Findings IC-001 Vision-language models perform near chance on the NL-Eye visual abductive reasoning benchmark IC-002 Even when VLMs select the correct hypothesis, their explanations are often invalid or unhelpful IC-007 Most LLMs do not align closely with human moral preferences on multilingual trolley problems IC-009 LLM moral preferences show significant language sensitivity but not inequality toward low-resource languages IC-025 LMMs exhibit poor fine-grained perception in locating individual characters on original oracle bones IC-026 LLMs can assist in OB rejoining by identifying rejoinable fragments with moderate accuracy, but are not yet truly usable IC-027 LMM performance in deciphering oracle bone inscriptions is comparable to untrained humans for common characters but declines for rarer and structurally complex characters IC-074 Released LLMs achieve F1 plan scores between 42.7 and 86.7 on the T-Eval plan task IC-078 GPT-4's self-verification loop causes performance collapse due to high false negative rates in binary verification IC-079 GPT-4's free-form critique generation is unreliable, containing hallucinated edges, vertex colors, and precondition states IC-080 GPT-4's performance is largely insensitive to the content of feedback; simple re-prompting with a sound verifier (sampling) matches or exceeds detailed critique IC-099 GPT-4 and Claude 3 Opus can be prompted to selectively underperform on WMDP while maintaining general performance on MMLU and CSQA IC-100 GPT-4, GPT-3.5, Claude 3, Llama 3 8B, and Llama 3 70B can be prompted to approximately calibrate their accuracy to specific target percentages on MMLU IC-101 GPT-4 and Claude 3 struggle to emulate a lower capability profile (high school freshman level) via zero-shot prompting, with only moderate improvement from chain-of-thought prompting IC-103 LLMs' value rankings align with the universal human value hierarchy under most prompting conditions IC-104 Value anchor prompting produces LLM value correlation structures that closely match the human circular value structure, while standard prompting does not IC-105 Value anchoring produces a sinusoidal scoring pattern around the circular value structure, with scores decreasing as circular distance from the anchor increases IC-1050 Released models generate patches that are less than half the length of gold solutions and rarely edit more than one file IC-109 GPT-3.5-turbo and GPT-4 produce cycles in inferred causal graphs when using pairwise prompts, with cycle counts growing sharply on larger graphs IC-110 Phi-3 (3.8B) and Llama-3 (8B) with triplet prompting outperform GPT-4 with pairwise prompting on causal graph orientation IC-1133 LLMs are highly receptive to coherent counter-memory as sole evidence, contradicting prior findings of stubbornness with entity-substitution counter-memory IC-1134 LLMs show strong confirmation bias in multi-source settings, preferring evidence consistent with parametric memory, with stronger bias for popular entities IC-1135 LLMs show order sensitivity to evidence position in context, with PaLM2 and LLaMA2-7B showing memorization ratio variations exceeding 30% IC-1144 GPT-4 and GPT-3.5 produce high rates of irrelevant (fabricated) books when answering constraint queries from parametric knowledge, with a sharp phase transition at low author popularity IC-1145 Providing complete context eliminates irrelevance but does not fix constraint satisfaction for GPT-4 or GPT-3.5 IC-1146 Self-context (self-retrieval) chain-of-thought increases the rate of fabricated books compared to no-context for both GPT-4 and GPT-3.5 IC-1147 GPT-4 outperforms GPT-3.5 on all KITAB metrics but the gap is modest, with all-correctness below 35% for both, suggesting scale alone does not resolve constraint satisfaction IC-1157 GPT-4 and other LMs show a large gap between rule induction and rule application, with task accuracy dropping to near zero on MiniScan when the LM itself applies its own proposed rules IC-1158 GPT-4 and other LMs are brittle to noisy exemplars and unfamiliar output representations, with performance degrading sharply even under minimal perturbation IC-1159 GPT-4 is a strong inductive hypothesis proposer, achieving high accuracy on inductive reasoning benchmarks when its generated rules are applied by a symbolic interpreter IC-1167 GPT-4 Code Interpreter's mathematical reasoning accuracy is positively correlated with code usage frequency, with its iterative code generation and self-debugging mechanism as the primary driver of its 69.69% zero-shot MATH accuracy IC-1168 For GPT-4 Code Interpreter, code-based self-verification improves MATH accuracy to 73.54% while natural language self-verification slightly degrades it to 69.29% relative to the 69.69% base IC-1170 GPT-3.5, Llama2, PaLM2, and GPT-4 are susceptible to a CoT-prompting backdoor attack (BadChain) on complex reasoning tasks, with stronger reasoning models showing higher attack success rates IC-1177 LLM-generated misinformation is harder for LLM detectors to detect than human-written misinformation with the same semantics IC-1188 GPT-3 procedural planning performance is scale-dependent, with Curie (6.7B) scoring 3.75 and Davinci (175B) scoring 4.90 overall quality in few-shot settings, while GPT-4 achieves 4.81 overall and 5.00 order IC-1193 Vicuna and Alpaca achieve 0% pass rate on all ToolBench tool-use instructions, while GPT-4 and ChatGPT reach 71.1% and 64.8% with DFSDT, revealing a wide capability gap in tool use among released LLMs IC-1194 GPT-3.5-turbo, GPT-4, and GPT-3.5-turbo-0613 exhibit 50-58% inconsistency between their ratings and rankings feedback on the same response pairs IC-1218 GPT-4 achieves 59.72% accuracy on choices13k and 80.3% on the horizon task when modeling human decisions IC-1220 GPT-4, GPT-3.5-turbo, and Llama-2-70B can implement learning algorithms in-context on novel boolean functions, competing with nearest-neighbor baselines IC-1225 GPT-4 generates realistic dynamic scene layouts from text prompts with only 3 in-context examples, achieving 98% average accuracy across 5 spatiotemporal tasks, with physics knowledge (gravity, elasticity, perspective) generalising to unseen objects from its weights IC-1240 GPT-4-0314 achieves 85% zero-shot accuracy on situational-awareness questions about its own architecture and training IC-1241 GPT-4 (14 March 2023) achieves 100% zero-shot accuracy at classifying whether news articles could be part of its pre-training data IC-1242 GPT-4 wrote a working script that called an instance of itself on its API as part of a plan to gain internet access IC-1243 GPT-4 and GPT-4 Turbo achieve near-zero scores on GAIA level 3 and single-digit to low-double-digit scores on levels 1-2, compared to 87-94% for human annotators IC-1244 GPT-4's non-zero scores on GAIA web browsing questions are largely due to memorization of intermediate information from training data rather than actual web browsing IC-1264 LLMs are overconfident when verbalizing confidence, with values concentrated in 80–100% and multiples of 5, yielding high ECE across all five tested models IC-1265 Calibration and failure prediction improve as model capability scales from GPT-3 to GPT-4, but remain far from ideal IC-1267 LLMs' alignment with human privacy judgments drops sharply as contextual complexity increases from tier 1 to tier 3 IC-1268 LLMs leak private information in theory-of-mind scenarios even when explicitly instructed to preserve privacy IC-1269 LLMs leak secrets to inappropriate recipients in meeting summarization and action-item generation tasks IC-1270 Chain-of-thought prompting does not mitigate privacy leakage in GPT-4 or ChatGPT IC-128 Released VLMs exhibit sycophancy, agreeing with incorrect user opinions while ignoring visual evidence, with LLaVA-1.5 showing the highest rate (94.6%) and InternLM-XComposer2-VL-1.8B the lowest (28.8%) IC-1288 GPT-4, ChatGPT, and GPT-4V fail to close the human-machine gap on Bongard-OpenWorld, with InstructBLIP captions differentially degrading ChatGPT while improving GPT-4 IC-1307 Released LLMs (CodeLlama 7B/13B/34B, GPT-3.5, GPT-4) achieve limited code-optimization speedups with standard prompting, with the best baseline (GPT-3.5 CoT) reaching only 1.60x versus the 3.66x human reference IC-1308 Dynamic retrieval-based few-shot prompting substantially improves released LLMs' code optimization, with GPT-4-0613 reaching 76.07% optimization rate and 3.93x speedup (best@8), exceeding the 3.66x human reference IC-1309 GPT-4-0613 exhibits reduced output diversity relative to GPT-3.5: it outperforms on best@1 but underperforms on best@8 under CoT prompting IC-131 GPT-4 Turbo, GPT-3.5 Turbo, Llama3-8B, Qwen-7B, and iFlytekSpark-13B over-rely on the strong reminder 'the answer is' in prompts as a shortcut, with accuracy dropping sharply when the cue is a random answer rather than the ground truth IC-1310 Chain-of-thought prompting provides notable code-optimization gains only for larger models (CodeLlama 34B, GPT-3.5, GPT-4) but not for CodeLlama 7B or 13B, consistent with an emergent capability IC-132 GPT-4 Turbo, GPT-3.5 Turbo, Llama3-8B, Qwen-7B, and iFlytekSpark-13B trust authority roles (teacher/judge) more than peer roles (classmate/lawyer) when the cue information is the correct answer IC-1327 GPT-4, GPT-3.5, Llama2, and Vicuna models underperform human annotators on multistep soft reasoning in natural language narratives, with smaller models scoring near random chance IC-1328 GPT-4's performance on MUSR depends on the type of neurosymbolic scaffolding: program-aided decomposition helps on structured optimization but symbolic belief tracking fails on natural language theory-of-mind IC-133 GPT-4, Claude 3, and Gemini 1.0 Pro do not exhibit detectable watermarks from the red-green, fixed-sampling, or cache-augmented families under black-box statistical tests IC-1330 All 20 evaluated LLMs improve in multi-turn task-solving with additional tool-use turns and GPT-4-simulated language feedback IC-1342 Assigning socio-demographic personas to LLMs causes significant reasoning performance degradation across all four models studied, manifesting as both explicit abstentions and implicit reasoning errors IC-1361 GPT-4 and other state-of-the-art LLMs achieve near-human accuracy in inferring personal attributes from unstructured text IC-1362 State-of-the-art text anonymization is insufficient to prevent GPT-4 from inferring personal attributes IC-1363 Current model alignment does not filter privacy-invasive prompts across major LLM providers IC-1364 GPT-4 can extract personal information from users through adversarial chatbot conversations IC-1373 All five evaluated LLMs show a strong positional bias in constrained text generation, with first-position constraints nearly always satisfied but last- and arbitrary-position constraints causing major performance drops IC-1374 Counting difficulty in constrained generation increases with text level and constraint strictness, with exact sentence-level character counts being the hardest condition for all five models IC-1375 GPT-4's constraint satisfaction improves by approximately 20% after one round of automated feedback but plateaus at 66% even after three additional rounds IC-1392 Title reproduction shows no contamination signal while tag reproduction shows a negative association with GitHub presence and a moderating difficulty effect for GPT-4 and GPT-3.5-turbo IC-1393 Most mainstream LLMs generate value-violating content at high rates (APV 65-80%) across 2,397 morally ambiguous prompts, indicating substantial ethical misalignment IC-1394 ChatGPT demonstrates better ethical value conformity than GPT-4 across multiple prompt generation sources IC-1401 GPT-4 serves as a proxy for human judgment on SOTOPIA-EVAL, with strong correlations on goal, financial, and relationship dimensions for model outputs IC-1403 All four evaluated LLMs produce negative scores on social rules and secret-keeping dimensions in SOTOPIA interactions IC-1404 On SOTOPIA-HARD, GPT-4 achieves significantly lower goal completion than humans and exhibits non-strategic negotiation and excessive compromise behaviors IC-1413 GPT-4 achieves near-saturation on Python code synthesis (86.6% pass@1) but scores significantly lower on code repair (47.8% avg) and code explanation (52.1% avg) across six languages IC-1458 GPT-4 and PALM 2-L produce significantly less consistent descriptions of interpolated domain embeddings than a purpose-built ELM model IC-1508 LLMs with in-context learning translate Kalamang-English at 44.7/45.8 CHRF, falling short of the human baseline of 51.6/57.0 CHRF IC-1509 Kalamang-English translation performance on MTOb increases with model size within the Llama and Llama 2 families, and GPT-4 outperforms Text-davinci-003 IC-1510 Without retrieved context, LLMs are unable to translate Kalamang, and among context types, retrieved parallel sentences are most beneficial, followed by word list entries, then grammar book passages IC-1523 Five AI assistants (Claude-1.3, Claude-2.0, GPT-3.5-turbo, GPT-4, Llama-2-70B-Chat) consistently exhibit sycophancy across four varied free-form text-generation tasks IC-1577 Base LLMs prompted with URiAL (3 restyled in-context examples + system prompt) match or surpass their SFT/RLHF-aligned counterparts on multi-aspect evaluation IC-1599 Self-repair at equivalent compute budget provides only modest and inconsistent gains over i.i.d. sampling for CodeLlama-13B-Instruct, GPT-3.5, and GPT-4 on HumanEval and APPS IC-1600 Replacing a model's self-generated feedback with a stronger model's feedback consistently improves self-repair beyond both the i.i.d. baseline and the self-repair baseline IC-1601 GPT-4's self-generated feedback is significantly less effective than human programmer feedback for code repair, with the gap widening on harder problems IC-163 LLaVA-1.5, LLaVA-Next, and GPT-4V show near-zero accuracy on GUI grounding benchmarks while achieving 50-85 on general image grounding (RefCOCO+), indicating a failure mode specific to GUI grounding scenarios IC-169 BT-based, DPO-based reward models, and GPT-4 as judge all exhibit significant length bias, with their scores correlating with output length rather than quality IC-170 GPT-4, GPT-3.5, and Claude-3.5-Sonnet rely heavily on parametric knowledge in RAG settings, producing ungrounded responses with high answered ratios and low trust-scores IC-171 ICL prompting produces binary response patterns in released LLMs, with answered ratios collapsing to near 0% or 100% rather than calibrated refusal, making prompting ineffective for RAG groundedness IC-188 LLMs show constraint-type-specific performance on system message following, with weaker models exhibiting large variance across constraint categories IC-189 Most LLMs show degraded instruction satisfaction when user instructions conflict with system messages, indicating difficulty in prioritizing system message constraints IC-190 LLMs show progressive degradation in system message constraint following across multi-turn conversations, with dependent conversations degrading faster than parallel ones IC-198 Safety-aligned LLMs (GPT-4, GPT-3.5, Gemma2-27b, GPT-4o, Gemma2-9b, Qwen2.5-72b, Mistral-7b, Mixtral-8x22b) are vulnerable to natural prompts semantically related to toxic seed prompts, with attack success rates of 82-99% IC-206 GPT-4o-0513 achieves the highest wb-reward mix score (35.7) on WildBench, with a clear three-tier structure among 40 evaluated LLMs IC-207 Open LLMs (Llama-3-8B-Inst, Yi-1.5-34B-Chat) show weaker performance on coding and math tasks compared to proprietary models (GPT-4-turbo-0409, Claude 3 Opus) which perform well across all task categories IC-230 Six LLMs show distinct value preferences on daily-life moral dilemmas, with significant inter-model differences on core values such as truthfulness and fairness IC-231 GPT-4-turbo and Claude-3-Haiku show inconsistent adherence to their providers' stated design principles when facing value conflicts in daily-life dilemmas IC-232 System prompts cannot effectively steer GPT-4-turbo's value preferences in moral dilemmas IC-252 GPT models produce harmful gender stereotypes at higher rates when user names imply a demographic group, with GPT-3.5 Turbo showing the highest rates and open-ended generation tasks most affected IC-253 Post-training reinforcement learning significantly reduces harmful gender stereotypes in GPT models, with the best-fit slope of 0.21 indicating post-RL models have far lower bias than pre-RL versions IC-257 327 DNNs approach or exceed human accuracy on object depth order but are near chance on VPT-basic, while humans show the opposite pattern IC-264 All 18 evaluated LLMs fail to abstain when the provided context lacks the answer, with performance gaps of 13.6% to 68.4% relative to the original context IC-265 Model families show extreme variation in detecting conflicting answers in inconsistent contexts, with phi-3 series at 5.8% average accuracy versus GPT-4 series at 89.35% IC-266 GPT-4o drops from 96.3% closed-book accuracy to 47.5% when given counterfactual context that contradicts its parametric knowledge, far below the 95% human accuracy on the same items IC-312 GPT-4o achieves the highest insight-level Llama-3-eval score (0.60) among all tested LLM backbones on InsightBench multi-step data analytics IC-313 All four LLM backbones fail to detect a planted linear trend in incident resolution time when the slope is below 0.1, and detection rates diverge sharply above that threshold IC-314 Llama-3-70b as an LLM-based evaluator (Llama-3-eval) produces agent rankings consistent with GPT-4-based G-Eval on InsightBench IC-331 In Stable Diffusion v1.5, specific tokens in text prompts drive memorization, and GPT-4-based perturbation of high-attribution tokens reduces SSIM similarity to training images while maintaining CLIP score IC-332 Llama-3-70B exhibits a friendlier, funnier, and less ethics-focused style than GPT-4 and Claude-3-Opus on Chatbot Arena, and these vibes predict model identity at 80% and user preference at 59% accuracy IC-334 GPT-4V produces more poetic, emotion-focused image captions compared to Gemini-1.5-Flash's literal descriptions, with 99% model-matching accuracy on COCO IC-335 GPT-4's detection performance as a scoring model is highly sensitive to the prompt, varying from 0.7289 to 0.9682 AUROC, far more than GPT-3.5 or Babbage IC-340 GCG jailbreaking attacks exhibit strong model-specific transferability, achieving below 3% ASR on Llama-2-13b-chat and Llama-3.1-8b-instruct but above 90% ASR on Vicuna-13b-v1.5 and Mistral-7b-instruct IC-351 GPT-4-turbo-2024-04-09, Qwen2-72b-instruct, and Llama-3.1-70b-instruct show distinct performance profiles across ultra-long, 32k, and 4k context benchmarks IC-352 RAG with sufficient retrieved tokens outperforms direct long-context for Qwen2-72b-instruct on >100k tasks, while at 32k the default RAG setting underperforms direct long-context for GPT-4-turbo-2024-04-09, Qwen2-72b-instruct, and Llama-3.1-70b-instruct IC-362 LLMs with chain-of-thought prompting predict and simulate human risky choices that are more rational than actual human behavior, correlating more highly with maximum expected value than with human choices IC-363 LLM inferences about others' preferences from observed decisions are highly correlated with human inferences because both assume the decision-maker is rational IC-383 GPT-4-1106, GPT-3.5-1106, and unfine-tuned CodeLlama-13b achieve 44.3%, 39.5%, and 38.5% API call accuracy respectively on unseen APIs (Level 3) with 3-shot retrieved prompting in the API Pack evaluation IC-388 CS-Bench scores correlate strongly (p > 0.9) with math and code benchmark scores across 12 models IC-389 All evaluated LLMs score significantly lower on CS reasoning questions than knowledge questions, with the gap narrowing for stronger models IC-392 Fuyu-8B and GPT-4V exhibit position bias in visual recognition, with model performance depending on where the target object appears in the image IC-393 GPT-4-turbo, Llama-3.1-8B-Instruct, and OpenAI Moderation show declining hate speech detection accuracy as sentence implicitness increases, with very low success rates in the highest implicitness ranges IC-407 Safety-aligned LLMs (Llama-2-chat, Llama-3-instruct, Gemma, GPT-3.5, GPT-4o, R2D2) achieve 100% jailbreak attack success rate under adaptive prompt-and-suffix attacks on 50 harmful requests IC-415 VLMs show a default shape bias (47.9-73.8%) that exceeds their vision encoders and vision-only models but falls short of human levels (96%), with the LLM component rather than the encoder responsible for suppressing one visual cue. IC-441 Skill-level improvements between model releases are highly uneven, with Claude 3.5 Sonnet gaining ~50% over Claude 3 Opus on law skills while Gemini improved most in math and science IC-474 GPT-4, GPT-4o, and Llama-3.1-405B fail at knowledge classification and comparison without chain-of-thought IC-475 GPT-4, GPT-3.5, GPT-4o, and Llama-3.1-405B fail at inverse knowledge search regardless of prompting IC-476 GPT-4 and GPT-3.5 show strong positional bias in Chinese idiom character completion IC-477 Released LLMs achieve limited success rates as web agents on WebArena-Lite, with open-source models substantially below proprietary ones IC-478 GPT-4 and GPT-4V achieve approximately 71-73% accuracy in judging whether a web agent trajectory successfully completes a task IC-485 LLMs show a significant performance gap between Wikipedia-based factual multi-hop QA and counterfactual multi-hop QA, indicating reliance on memorized knowledge rather than reasoning from context IC-486 LLMs achieve correct final answers through incorrect reasoning chains, inflating their apparent multi-step reasoning performance IC-487 Including sub-questions in the prompt improves LLM performance on multi-hop QA tasks IC-488 LLM performance degrades progressively as the number of reasoning hops increases, with error propagation from earlier sub-questions IC-495 All evaluated multimodal foundation models achieve average non-hallucination accuracy below 50% across six hallucination scenarios IC-496 GPT-4o achieves the highest location inference accuracy among evaluated models, reaching 98.16% for country, 60.23% for city, and 27.13% for zip code from street view images IC-497 Text-to-image models experience performance drops exceeding 10% under adversarial prompts, with spatial reasoning being the most vulnerable task across all models IC-498 Multimodal foundation models exhibit severe group unfairness, with race and age biases more pronounced than gender bias in text-to-image models while gender bias is stronger in image-to-text models IC-508 GPT-4 exhibits reduced preference consistency (0.66 vs 0.84) when the quality distinction between two responses is minimal IC-509 GPT-4 used as a preference labeler via prompt engineering yields alignment performance comparable to a task-specific 125M model IC-518 GPT-4o achieves 62.54% overall accuracy on MMWorld, the best among 15 MLLMs, while four open-source models perform below the 26.31% random-choice baseline IC-519 MLLMs exhibit different skill sets than humans, correctly answering expert-level questions that all three human annotators miss while failing on easy questions humans answer correctly IC-520 Temporal reasoning performance drops significantly across all 15 MLLMs when video frames are shuffled or reduced to one-fifth of the original count IC-522 LLMs perform correct example inference without inducing the correct rule, and this gap is robust to prompting methods, fact count, and scenario form IC-527 A 16 million latent sparse autoencoder substituted into GPT-4 yields a language modeling loss corresponding to 10% of GPT-4's pretraining compute IC-549 All 18 evaluated LLMs show a 15-20% performance gap between linear (node chain) and graph (workflow) planning on WorfBench IC-551 GPT-4's workflow generation performance declines as the number of nodes and edges in the workflow increases IC-552 GPT-4, Llama-3.1-8B, and Qwen-2-72B all improve on ALFWorld and WebShop when given a generated workflow as structured prior knowledge IC-596 Multiple released LLMs fail to refuse harmful prompts disguised as historical or philosophical discussions, with GPT-4o and Mixtral showing the lowest refusal rates IC-682 GPT-3.5 and GPT-4 achieve F1 scores of 0.5820 and 0.6180 respectively on pairwise response quality evaluation against human annotations IC-687 GPT-3.5+ models exhibit a gambler's fallacy bias and generate low-complexity sequences when asked to produce random binary sequences IC-689 Subjective randomness generation and sharp ICL transitions emerge only in larger or reward-fine-tuned models, absent in earlier GPT-3 variants and smaller open-source models IC-700 GPT-4's reasoning accuracy degrades when provided with incorrect hints from a weaker model IC-714 For GPT-4 book-length summaries, human annotators prefer incremental summaries for detail (83% vs 11%) but hierarchical for structure (59% vs 35%), logic (53% vs 38%), and overall (54% vs 44%), showing coherence and human preference are not aligned IC-718 GPT-4 achieves 0.882 Pearson correlation with human evaluators on 45 customized score rubrics while GPT-3.5-Turbo achieves only 0.392 IC-719 Llama-2-Chat achieves reasonable human-preference accuracy (51.78-53.67%) as a prompted reward model without specific reward-model training IC-739 GPT-3.5-turbo-0613 combined with CLIP produces more faithful concept-salience pseudo-labels than LLaMA-2-13B-Chat, InstructBLIP, or LLaVA-1.5B on most of five datasets IC-750 Open-source models without safety training are significantly more vulnerable to jailbreak attacks than safety-aligned proprietary models IC-751 Arena-Hard-200 reveals larger performance gaps between open and proprietary LLMs than MT-Bench IC-752 GPT-4's win rate over GPT-3.5-turbo is 52% on the top-50 most challenging prompts but only 22% on the bottom-50 easiest prompts IC-761 GPT-3.5-turbo and GPT-4 are susceptible to specific circulating jailbreaking prompts, with 'jailmommy' achieving a 71.16% success rate in producing toxic outputs IC-762 GPT-4 and GPT-3.5 outperform humans in generation but underperform in discriminative (selective) evaluation across 10 of 13 language tasks IC-764 GPT-4 and GPT-3.5 make frequent errors answering questions about their own generated text, underperforming humans in interrogative evaluation IC-790 GPT-4 outperforms GPT-3.5-turbo, Vicuna-13B, and Llama2-70B for generating target captions in zero-shot compositional image retrieval IC-798 GPT-4 achieves 61.6% on MATH with tool-integrated reasoning prompting, exceeding PaL (51.8%) and CoT (42.5%) IC-798 GPT-4 achieves 61.6% on MATH with tool-integrated reasoning prompting, exceeding PaL (51.8%) and CoT (42.5%) [context] IC-827 LLMs exhibit distinct psychological profiles that differ from human norms and vary by model size and version IC-828 Jailbreaking GPT-4 via cipherchat shifts its psychological profile toward human norms and reduces emotional intelligence scores IC-852 Intrinsic self-correction without external feedback consistently degrades reasoning accuracy across GPT-3.5-turbo, GPT-4, GPT-4-turbo, and LLaMA-2-70B-chat IC-868 Skill-Mix performance degrades with increasing k, and within the Llama-2 family the saturation point increases with model size IC-870 GPT-4's performance on Skill-Mix(k=5) and Skill-Mix(k=6) provides probabilistic evidence of generating novel skill-topic combinations not present in training data IC-871 Llama-2-70b-chat as a grader is more generous than GPT-4 and systematically gives higher scores to Llama-2 family outputs IC-897 GPT-4 underperforms codex on Spider text-to-SQL with few-shot prompting, attributed to its zero-shot tuning IC-898 GPT-3.5-turbo and GPT-4 are overconfident in their initial code predictions when unit test execution is unavailable IC-899 GPT-4, when used as a blind pairwise evaluator, exhibits the same style-over-factuality preference as human crowdworkers IC-927 Human ciphers (ASCII, Unicode, Caesar, Morse) bypass the safety alignment of GPT-4 and GPT-3.5-turbo, with more powerful models producing more unsafe responses IC-928 SelfCipher (a role-play prompt without explicit cipher rules) evokes a 'secret cipher' in LLMs, achieving high unsafety rates that outperform most human ciphers IC-929 Simulated character-level ciphers that never appear in pretraining data cannot bypass safety alignment even with 10+ demonstrations IC-931 GPT-4 achieves approximately 95% accuracy on contrast instructions, far exceeding human performance without tools Shared mechanisms Circular representation also in Gemini 1.0 Pro , Gemma 2 , GPT-2 , Llama 3 , Llama 3.1 , Mistral 7B / Mistral / Mistral 3 7B / Mistral-0.2-7B / Mistral-v0.1 , TerraMind Explanation faithfulness also in BakLLaVA , CF2 , Claude 3 , Claude 3.5 , CLIP / CLIP-ViT (LC) , DRUM , Fuyu , GEM , Gemini 1.5 / Gemini Pro 1.5 , Gemma 2 , GPT-3.5 / ChatGPT-3.5 , GPT-4o , GPT-J , Idefics , Llama 3.1 , LLaVA-NeXT / LLaVA 1.6 , MAE-B/16 , MobileNetV2 , mPLUG-Owl3 , OpenFlamingo , PGExplainer , Pythia , RCExplainer , ResNet / ResNet-152 / ResNet-101 / ResNet-50-BN , SigLIP , SigLIP-2 , Stable Diffusion , TAGExplainer , ViT Failure mode also in AASIST , ADM , Aegis-Guard-Defensive , Alpaca , AnyLoc , AutoTikZ / DataTikZ , Baichuan , Baichuan 2 , Baichuan2-13B , BakLLaVA , Bard , BEiT , BERT , BingChat , BLIP , BLIP-2 , BLOOM , BSRNet , CF2 , Chat-UniVi-7B , ChatGLM-6B / ChatGLM-6b-2 , ChatGLM2 , ChatGPT , CLAP , Claude 1.3 , Claude 2.0 , Claude 2.1 , Claude 3 , Claude 3.5 , CLEAR , CLIP / CLIP-ViT (LC) , CLIP4Clip , CLIPBERT , CLIPCap , CLMBR-T-BASE , CloFNet , Code Llama , CodeGeex2 , CodeGen , CodeLlama-13B , CodeLlama-34B , CogVLM2 , Cohere Command R , CoMEt , Command R+ , CONCH , CycleGAN , DALL-E , DALL·E 2 , DALL·E 3 , DASR , DECAF , DeepSeek-2-Chat , DeepSeek-2-Coder , DeepSeek-V2-0628 , DeepSeek-VL , DeepSeek-VL2 , DeiT , DeiT-III , Depth Anything , DETR-R50 , DimeNet++ , DINO , DINO-FocalNet-Large , DINOv2 , EGNN , Emu2 , EquiformerV2 , ESCN , ESM-2 , ESM3 , ESRGAN , EVA-CLIP , EVE , Falcon , Faster R-CNN / Faster R-CNN R50 / Faster R-CNN X101 , FLAN-T5 , Florence-2 , FLUX / FLUX1 , FoveaBox , Fuyu , Galactica-6.7B , GAT , 3D Gaussian Splatting , GCN , Gemini , Gemini 1.5 / Gemini Pro 1.5 , Gemma , Gemma 2 , GIN , GLIDE , GLM-4 , GLM-4V , GloVe , GP-UNIT , GPT-2 , GPT-3 / GPT base , GPT-3.5 / ChatGPT-3.5 , GPT-4.1 , GPT-4o , GPT-J , GPT-NeoX-20B , GraphSAGE , Grounding DINO , Guanaco , GVP , Hawkeye , HiFaceGAN , HPS , HuggingGPT , IDDPM , Idefics , Idefics2 , ImageBind , ImageBind-LLM-7B , Imagen Video , ImageReward , 12-in-1 , InstructBLIP , InstructGPT , InternLM-2.5-7B , InternLM-XComposer2-VL , InternVideo , InternVL-1.5 , InternVL2 , Koala , LegalBERT , LLaMA , Llama 2 / Llama 2 base , Llama 3 , Llama 3.1 , Llama 3.2 , Llama-3.2-3B , Llama-3-2-Vision , LLaMA-Adapter v2 , Llama Guard , Llama Guard 2 , Llama-Guard 3 , Llama-VID , LLaVA , LLaVA-1.5 / LLaVA-v1.5 , LLaVA-Med , LLaVA-NeXT / LLaVA 1.6 , LLaVA-OneVision , LongVA-7B , LOVT , LWM-1M-JAX , MACE , MAE , Med-Flamingo , Merlot Reserve , MGCA , MiDaS , MiniCPM-V , MiniGPT-4 , Mip-Splatting , Mistral 7B / Mistral / Mistral 3 7B / Mistral-0.2-7B / Mistral-v0.1 , Mistral Large 2 , Mistral Large V2 , Mixtral , Mixtral 8x7B / Mistral 8x7B Instruct / Mixtral 46.7B / Mixtral 8x7B Instruct / Mixtral-instruct-8x7b , MobileNetV2 , Molmo , MolmoE-7B , Momentor , Moondream2 , Moonshot-v1-8k , mPLUG-2 , mPLUG-Owl , mPLUG-Owl3 , mPLUG-Owl2 , MPT , MSA Transformer , MultiBERTs , Nova Canvas , Nova Lite , Nova Pro , O1 / OpenAI-o1-preview , O3 , O4-mini , OLMo / OLMo base , OneLLM , OpenAI Moderation , OpenChat-3.5-0106 , OpenCLIP , OpenFlamingo , OPT , Otter , Otter-7B , PaLM 2 , PaLM 62B , PandaGPT-7B , PerSAM , Phi-3 , Phi-3.5 Mini Instruct , PickScore , PLIP , Prismatic , ProGen-2 , Pythia , Qwen1.5 , Qwen 2 , Qwen 2.5 72B Instruct , Qwen2-VL , Qwen-Audio , Qwen-VL , Qwen2.5 , Qwen2-Audio , R2D2 , RadFM , RCExplainer , RD-SR , RealESRNet , Reprover , ResNet / ResNet-152 / ResNet-101 / ResNet-50-BN , RetinaNet-R50 , RivaGAN , RS-LDS , RT-1 , RT-2 , SALMONN , SAM , SAM 2 , SAULLM 54B , Scaffold-GS , SchNet , Seed-LLaMA-8B , SGC , SIREN , Sketch Transformer , SLD-max , SLD-medium , SLD-strong , SLDS , SLIMG , SpeechGPT , SphereNet , SRResNet , Stable Diffusion , StackLLaMA , Starcoder , StegaStamp , StyleGAN2-ADA , Swin Transformer , T5 , TD-MPC , TerraMind , TimeChat , TranceptionEVE , TreeRing , Tulu 2 , UnifiedQA , UniPerceiver , UNITER , UniVL , Van Gogh Diffusion , VERA , VGG / VGG13 , Vicuna , Video-Chat-7B , Video-ChatGPT , Video-LLaMA , Video-LLaMA-2-13B , Video-LLaVA , VideoCLIP , ViLA-8B , ViLBERT , VindLU , VioLET , ViRTex , ViT , ViV1T , VTG-LLM , WildGuard , Wizardlm , X-CLIP , X-InstructBLIP-7B , XGen-MM , Xlm-R , Zephyr-7B-beta Positional bias also in BERT , ChatGPT , Claude 3 , Claude 3.5 , Falcon , Fuyu , Gemini , Gemini 1.5 / Gemini Pro 1.5 , Gemma , Gemma 2 , GLIDE , GPT-2 , GPT-3.5 / ChatGPT-3.5 , GPT-4.1 , GPT-4o , GPT-J , InstructGPT , LLaMA , Llama 2 / Llama 2 base , Llama 3 , Mistral 7B / Mistral / Mistral 3 7B / Mistral-0.2-7B / Mistral-v0.1 , Mixtral , Mixtral 8x7B / Mistral 8x7B Instruct / Mixtral 46.7B / Mixtral 8x7B Instruct / Mixtral-instruct-8x7b , MPT , O1 / OpenAI-o1-preview , O3 , O4-mini , PaLM 2 , Phi-3 , Pythia , Qwen1.5 , Qwen 2 , Qwen 2.5 72B Instruct , Stable Diffusion , Sybil , Vicuna Scale-dependent behaviour also in Aquila-7B , BEiT , BERT , BLOOM , Claude 2.1 , Claude 3 , Claude 3.5 , CLIP / CLIP-ViT (LC) , Code Llama , CodeGen , Cohere Command R , DeepSeek LLM , DeepSeekMoE , DeiT-III , DINO , DINOv2 , EquiformerV2 , ESCN , Falcon , FLAN-T5 , Gemini 1.0 Pro , Gemini 1.5 / Gemini Pro 1.5 , Gemma , Gemma 2 , GPT-2 , GPT-3 / GPT base , GPT-3.5 / ChatGPT-3.5 , GPT-4o , GPT-J , GPT-Neo , I3D , Idefics , InternLM-2.5-7B , InternLM-XComposer2-VL , InternLM2 , InternVL-1.5 , InternVL2 , LLaMA , Llama 2 / Llama 2 base , Llama 3 , Llama 3.1 , Llama 3.2 , Llama-3.2-3B , LLaVA-1.5 / LLaVA-v1.5 , LLaVA-NeXT / LLaVA 1.6 , LongVA-7B , MAE , Mistral 7B / Mistral / Mistral 3 7B / Mistral-0.2-7B / Mistral-v0.1 , Mixtral , Mixtral 8x7B / Mistral 8x7B Instruct / Mixtral 46.7B / Mixtral 8x7B Instruct / Mixtral-instruct-8x7b , Moirai , MPT , MViT V2 , O1 / OpenAI-o1-preview , OLMo / OLMo base , OpenCLIP , OpenFlamingo , OpenLLaMA , OPT , PaLM 2 , Phi-3 , Platypus2-Instruct-70B , Pythia , Qwen , Qwen1.5 , Qwen 2 , Qwen 2.5 72B Instruct , Qwen2-VL , Qwen-Audio , Qwen2.5 , RedPajama-INCITE , ResNet / ResNet-152 / ResNet-101 / ResNet-50-BN , SlowFast , Solar 10.7B , Stable Diffusion , StableLM , Swin Transformer , T5 , TerraMind , text-ada-001 , TigerBot-70B-Chat , TimesFormer , TSM , Tulu 2 , Uniformer , Vicuna , VideoMAE , ViLA-8B , Wizardlm , X3D , XGLM , Xwin-LM-70B-v0.1 , Yi Shortcut also in BakLLaVA , BLIP-2 , Claude 3 , Claude 3.5 , CLIP / CLIP-ViT (LC) , CLIP4Clip , CLIPBERT , DALL·E 2 , DALL·E 3 , DeepSeek-VL2 , Eurus-RM-7B , Falcon , FLUX / FLUX1 , Gemini , Gemini 1.5 / Gemini Pro 1.5 , Gemma , Gemma 2 , GPT-2 , GPT-3 / GPT base , GPT-3.5 / ChatGPT-3.5 , GPT-4o , GPT-J , iFlytekSpark-13B , InstructBLIP , Internlm2-Reward , InternVideo , InternVL2 , LLaMA , Llama 2 / Llama 2 base , Llama 3 , Llama-3-2-Vision , LLaVA , LLaVA-1.5 / LLaVA-v1.5 , LLaVA-Med , LLaVA-NeXT / LLaVA 1.6 , Med-Flamingo , Merlot Reserve , MiniGPT-4 , Mistral 7B / Mistral / Mistral 3 7B / Mistral-0.2-7B / Mistral-v0.1 , Mixtral , Mixtral 8x7B / Mistral 8x7B Instruct / Mixtral 46.7B / Mixtral 8x7B Instruct / Mixtral-instruct-8x7b , Molmo , mPLUG-2 , mPLUG-Owl3 , Nova Canvas , O1 / OpenAI-o1-preview , OpenCLIP , OPT , Otter , PaLM 62B , Pythia , Qwen , Qwen 2 , Qwen-VL , Qwen2.5 , RadFM , ResNet / ResNet-152 / ResNet-101 / ResNet-50-BN , Stable Diffusion , Swin Transformer , Sybil , TerraMind , Tulu 2 , UniPerceiver , UniVL , Vicuna , Video-LLaMA , VideoCLIP , VindLU , VioLET , X-CLIP