An LLM's core training objective—predicting the next token—makes it sensitive to the raw frequency of words and numbers online. This creates a subtle but profound flaw: it's more likely to output '30' than '29' in a counting task, not because of logic, but because '30' is statistically more common in its training data.

Contamination in coding benchmarks is subtle. Instead of just spitting out a known solution, models like GPT-5.2 use implicit knowledge from their training data (e.g., popular codebases) to reason about unstated requirements. This makes it hard to distinguish true capability from memorization, as the model's 'chain of thought' appears logical while relying on leaked information.

If a document in the validation set has a near-duplicate in the training set, the model's high score is a lie. It's not demonstrating generalization; it's recalling something it has already seen. To prevent this, teams must deduplicate the entire dataset before splitting it into train and validation sets.

MIT research reveals that large language models develop "spurious correlations" by associating sentence patterns with topics. This cognitive shortcut causes them to give domain-appropriate answers to nonsensical queries if the grammatical structure is familiar, bypassing logical analysis of the actual words.

Simply stuffing all historical data into a large context window is counterproductive. The model's attention gets diluted by repetitive tool logs and intermediate data, making it struggle to find original instructions. This "signal versus noise" problem leads to hallucinations and degraded performance.

Newer LLMs exhibit a more homogenized writing style than earlier versions like GPT-3. This is due to "style burn-in," where training on outputs from previous generations reinforces a specific, often less creative, tone. The model’s style becomes path-dependent, losing the raw variety of its original training data.

Contrary to the belief that memorization requires multiple training epochs, large language models demonstrate the capacity to perfectly recall specific information after seeing it only once. This surprising phenomenon highlights how understudied the information theory behind LLMs still is.

The idea of separating "fact learning" from "skill learning" is a false dichotomy. Models need a base of internalized facts to reason effectively. The key is developing intelligence to compress what's important and discard what isn't, much like lossy human memory.

Research from Google shows that repeating key messages within a prompt's context window improves an LLM's recall and assigns more weight to that information. This suggests that future SEO for AI will involve strategic repetition, not just unique content.