DENIAHL, a new synthetic benchmark, shows that LLM needle-in-a-haystack recall is influenced by data type, item length, and pattern structure, not just context length.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DENIAHL: In-Context Features Influence LLM Needle-In-A-Haystack Abilities
DENIAHL, a new synthetic benchmark, shows that LLM needle-in-a-haystack recall is influenced by data type, item length, and pattern structure, not just context length.