SALMONN: dual audio encoders with activation-tuned language understanding

What changed

Integrated speech and general-audio encoders with a pretrained language model and used few-shot activation tuning to elicit broader audio-conditioned behavior.

Scientific scope

Integrated speech and general-audio encoders with a pretrained language model and used few-shot activation tuning to elicit broader audio-conditioned behavior. Inherits pretrained encoders/LLMs and cross-modal adapters. Emergence/first/generic-hearing claims are source-specific, not proof of universal hearing or unrestricted unseen-task transfer.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits pretrained encoders/LLMs and cross-modal adapters. Emergence/first/generic-hearing claims are source-specific, not proof of universal hearing or unrestricted unseen-task transfer.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• SALMONN: Towards Generic Hearing Abilities for Large Language Models

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction