Sometimes. A language model can apply a pattern to examples it has not seen, particularly when the examples reveal how familiar parts fit together. But getting a new example right does not, by itself, prove the model learned a general rule: results depend on the task, the demonstrations, and what counts as “new.”
What would count as learning the rule?
Consider this puzzle: a made-up operation, blick, turns “A then B” into “B then A.” If a model is shown that blick(red, blue) produces “blue then red,” can it correctly handle blick(circle, square)? That answer would be more informative than repeating the first example—but it would still leave questions. Were the new items familiar in other contexts? Did the prompt show enough examples to suggest the transformation? Would the model succeed with a longer sequence or a different kind of input?
In this context, compositional generalization means handling a new combination of parts the model has encountered before. In-context learning means responding to examples included in a prompt, without fine-tuning the model for that particular task. These behaviors can look like rule use, but an answer alone cannot tell us whether the model formed a symbolic rule, combined previously learned skills, or used some other learned mechanism.
That distinction matters because success on familiar-looking examples can arise without reliable transfer. A stronger test holds out cases that require applying the apparent rule in a genuinely new way—and states exactly what was held out. In a 2025 PNAS study of hidden-rule and symbolic tasks, Song, Xu, and Zhong describe out-of-distribution generalization as an ability that can appear with appropriately formatted prompts, while noting that its underlying mechanisms remain poorly understood: “Out-of-distribution generalization via composition: A lens through induction heads in Transformers”.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
- Read Before You Buy — No Video Output: These adapters support charging and USB 2.0 data transfer, but cannot transmit video signals. Except for standard USB webcams (which use USB data only), they are not compatible with HDMI/DisplayPort cables, video-capable USB-C hubs, or docking stations with video output.
- Convert USB-A Ports to USB-C: Designed to connect USB-C earphones, cables, flash drives, card readers, and other USB-C accessories to standard USB-A ports. Plug-and-play with no drivers or software required.
- Aluminum Alloy Housing: Built with a sturdy aluminum alloy shell that aids in heat dissipation and protects against daily wear and scratches. Designed to maintain a stable and secure connection.
- Compact & Travel-Friendly: The ultra-compact design allows the adapter to stay plugged into your device without blocking adjacent ports or adding bulk, reducing wear and tear on your original USB ports.
- 12-Month Warranty: Backed by a 12-month manufacturer warranty for peace of mind. Designed to meet strict quality control standards for reliable everyday performance.
What the studies show—and where they stop
The findings are conditional, not a yes-or-no verdict about all language models. Each study tests a particular kind of generalization, using particular prompts, models, and evaluation sets. Its result should be read in that context.
| Study | What was tested | What the result supports | What it does not establish |
|---|---|---|---|
| Song, Xu, and Zhong, PNAS (2025) | Hidden-rule and symbolic reasoning tasks. | Compositional structure can help explain out-of-distribution generalization in the settings studied. | A reliable universal rule-learning mechanism; the authors say the underlying mechanisms remain poorly understood. Source |
| Chen et al., Findings of EMNLP (2024) | A “Skills-in-Context” prompt format across the tasks tested. | Prompts showing foundational skills as well as examples that compose them can elicit systematic generalization. The authors report that some tasks needed as few as two exemplars. | That two examples—or this prompt format—will work for any task. “Near-perfect” describes the authors’ results on their tested tasks, not a general guarantee. Source |
| An et al., ACL (2023) | How the selected in-context examples affect compositional generalization. | Example selection matters: their experiments favor examples that are structurally similar to the test case, diverse from one another, and individually simple. Coverage of the linguistic structures needed for the test also matters. | That any one selection recipe ensures success; they also report weaker generalization on fictional words. Source |
| Lake and Baroni, Nature (2023) | A meta-learning model on SCAN systematic-generalization splits and other structural tasks. | The model achieved at least 99.78% accuracy on three SCAN lexical-generalization splits, showing strong performance in those specific conditions. | That the same performance transfers to other kinds of generalization: the study reports failures on other structural splits. Source |
| Mészáros et al., NeurIPS (2024) | Formal-language prompts for “rule extrapolation,” where a prompt violates at least one rule. | Evaluations can test a distinct and demanding case: whether a model handles prompts that break a formal rule. | That this test is interchangeable with learning a new combination of familiar parts. The study’s definition makes clear that the precise change between examples and test cases matters. Source |
| Hosseini et al., BlackboxNLP (2022) | Compositional generalization across three semantic-parsing datasets and four model families. | The authors report a decreasing relative generalization gap with scale in those evaluations. | That scaling eliminates compositional limits or predicts performance on unrelated tasks. Source |
The numbers in these studies are benchmark-specific experimental results. They do not provide a single rate for how often language models learn rules across tasks or in everyday use.
Rank #2
- 5-in-1 USB-C Hub: Experience comprehensive connectivity featuring a Power Delivery input, two USB-A 2.0 ports, a USB-A 3.0 port, and an HDMI port. (Note: The USB-C power delivery input port is only for connecting an external wall charger to power your laptop and cannot power peripheral devices.)
- 90W Pass-Through Charging: Achieve optimal charging with 90W pass-through power to your laptop, supported by a total input of 100W, with the hub reserving 10W for operational efficiency. (Note: Wall charger not included.)
- Quick Data Transfers: Accelerate your productivity with rapid data transfers using a high-speed 5Gbps USB 3.0 port and two 480Mbps USB 2.0 ports.
- 4K HDMI Display: Enhance your visual experience with a hub capable of delivering 4K resolution at 30Hz in both mirror and extend modes. Please note that this hub is compatible with MacBook (macOS 12 and newer), Windows 10 and 11, ChromeOS, and laptops equipped with DP Alt Mode and Power Delivery. Note: This device is not compatible with Linux.
- What You Get: Anker USB-C Hub (5-in-1, 4K HDMI), welcome guide, 18-month warranty, and our friendly customer service.
Why can a model succeed on one new example and fail on another?
The test may change more than one thing
A model might handle a new combination of familiar words but fail when asked to apply a rule to a longer sequence, a new sentence structure, or a prompt that violates a formal rule. Those tests place different demands on the model. A result on one does not settle the others.
The examples may not cover the needed structure
A prompt can contain many demonstrations yet omit the particular relationship needed for the test. In An et al.’s experiments, structurally similar, varied, and simple examples were more helpful than example selection without regard to the test structure. This is a reason to evaluate both the apparent rule and whether the demonstrations actually expose the structure it requires.
Rank #3
- Sleek 7-in-1 USB-C Hub: Features an HDMI port, two USB-A 3.0 ports, and a USB-C data port, each providing 5Gbps transfer speeds. It also includes a USB-C PD input port for charging up to 100W and dual SD and TF card slots, all in a compact design.
- Flawless 4K@60Hz Video with HDMI: Delivers exceptional clarity and smoothness with its 4K@60Hz HDMI port, making it ideal for high-definition presentations and entertainment. (Note: Only the HDMI port supports video projection; the USB-C port is for data transfer only.)
- Double Up on Efficiency: The two USB-A 3.0 ports and a USB-C port support a fast 5Gbps data rate, significantly boosting your transfer speeds and improving productivity.
- Fast and Reliable 85W Charging: Offers high-capacity, speedy charging for laptops up to 85W, so you spend less time tethered to an outlet and more time being productive.
- What You Get: Anker USB-C Hub (7-in-1), welcome guide, 18-month warranty, and our friendly customer service.
Familiar words and invented symbols are not equivalent
Ordinary words may come with patterns learned during pretraining. A model could therefore benefit from prior familiarity even when a prompt looks like a new rule-learning task. An et al. report weaker in-context generalization on fictional words, underscoring why tests with familiar language and unfamiliar symbols can yield different results.
Success on a benchmark may be narrow
Lake and Baroni’s meta-learning model did very well on three SCAN lexical-generalization splits, but the same study found failures on other structural splits. As the authors put it, “Systematicity continues to challenge models.” Strong performance is meaningful evidence for the tested case; it is not proof of broad, dependable generalization.
Rank #4
- Dual Converters, Infinite Potential：Includes 2× USB C male to USB A female adapters and 2× USB A male to USB C female adapters. Perfect for a wide range of uses—tablets with Bluetooth keyboards, expand USB ports on macbook, and more. Two different converters for all your daily needs
- Next-Level 10Gbps & 3A Charging: No more slow 480Mbps, this usb to usb c adapter has a transfer speed of up to 10Gbps, allowing you to do more transferring in less time. This usb adapter fits both USB A and USB C charger, supporting up to 3A fast charging
- Upgraded Exquisite Craftsmanship: With an aluminum alloy housing and metal connector, the usbc to usb adapter is extremely durable and sturdy. Rigorously tested to withstand more than 10,000 times of plugging and unplugging, ensuring long-lasting performance
- Broad Compatible: The usb c to usb adapter widely supports all USB C/ USB A devices like laptops, tablets, cellphones, car chargers, and phone chargers. Such as compatible with MacBook Pro/Air 2023/2022, Thunderbolt 4/3 Devices,Apple MagSafe Watch 9/8/7/SE/Ultra, iPad Pro 2022/2021, Samsung Galaxy S23/S20/S10, and iPhone 17/16/15 Pro. Plug and play
- Please Note: To reach 10Gbps speed, keep the cable under 3.3 ft. For USB A Male to USB C adapters, try flipping the USB C connector. USB C Male to USB A adapters support bidirectional 10Gbps transfer within 3.3 ft
How to tell whether an evaluation tests rule-like generalization
When judging a claim that a model learned a pattern, look for the specific difference between demonstrations and test cases. A useful evaluation describes what stayed familiar and what changed:
- New combination of known parts: Are the pieces familiar, with only their combination held out?
- New words or symbols: Does the model have to apply the pattern when familiar language cues are removed?
- Greater structural complexity: Does it handle a longer sequence or a new sentence structure, rather than only a short rearrangement?
- Rule-violating prompt: Is the test asking the model to extrapolate when a prompt itself breaks a formal rule, as in the rule-extrapolation setup studied by Mészáros et al.?
- Demonstration coverage: Do the examples reveal the component skills and the way they must be combined, or merely show unrelated input-output pairs?
- Held-out cases: Were the evaluation examples genuinely excluded from the demonstrations, and is the type of holdout clearly stated?
These distinctions make reported scores easier to interpret. A model’s success on familiar vocabulary, a new combination of known components, and a rule-violating formal-language prompt are not interchangeable forms of evidence.
Best Value
- 5-in-1 Connectivity: Equipped with a 4K HDMI port, a 5 Gbps USB-C data port, two 5 Gbps USB-A ports, and a USB C 100W PD-IN port. Note: The USB C 100W PD-IN port supports only charging and does not support data transfer devices such as headphones or speakers.
- Powerful Pass-Through Charging: Supports up to 85W pass-through charging so you can power up your laptop while you use the hub. Note: Pass-through charging requires a charger (not included). Note: To achieve full power for iPad, we recommend using a 45W wall charger.
- Transfer Files in Seconds: Move files to and from your laptop at speeds of up to 5 Gbps via the USB-C and USB-A data ports. Note: The USB C 5Gbps Data port does not support video output.
- HD Display: Connect to the HDMI port to stream or mirror content to an external monitor in resolutions of up to 4K@30Hz. Note: The USB-C ports do not support video output.
- What You Get: Anker 332 USB-C Hub (5-in-1), welcome guide, our worry-free 18-month warranty, and friendly customer service.
So, can a language model learn the rule behind a pattern?
It can sometimes behave as though it has learned a rule, by applying a pattern to cases not shown in the prompt. Carefully designed demonstrations can help, and some evaluated systems generalize impressively on specific tasks. But performance varies with the rule, the examples, the symbols, and the kind of novelty being tested. The cited findings support rule-like generalization in some settings; they do not establish that models reliably discover one general-purpose rule behind any pattern, or that their internal process is the same as a person’s.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
