PI-LLM

A key-value overwrite test of proactive interference: the model must report each key's last value after many similar updates, isolating working-memory limits inside the context window.

Also known as: PI-LLM Bench, PI_LLM, pi-llm

unassessed

This page is a discovery lead. Nobody has yet assessed it against the catalogue contract, so it carries no disposition. Absence of evidence here is not evidence of staleness.
Categorylong-context
Subcategoryproactive interference / key-value working-memory retrieval
Page statusactive
Metricauc_log1.5 (primary); average accuracy as a reference
Directionhigher_is_better
Dataset size740
Dataset licenceMIT
PublisherUniversity of Virginia (Physics) and NYU Center for Neuroscience

What it measures

PI-LLM (Proactive Interference in LLMs) streams many semantically related key-value updates and then asks only for the current, last value of each key. The items sit well inside typical context windows (the OpenCompass README gives a 5-25k token band). Accuracy still falls as earlier values overwrite later ones, which the authors treat as working-memory interference rather than a haystack-length problem. Four experiment types vary update count (2-400), concurrent keys, value length (1-40 characters), and sequential versus randomised update order.

Task format

English text generation over a chat-style prompt of key-value lines plus a final question asking for the last value of one or more keys. OpenCompass runs zero-shot generation (max_out_len 2048) and grades the free-text reply against a JSON ground-truth value.

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub