Search papers, labs, and topics across Lattice.
6
0
9
State-of-the-art VLMs struggle to execute scientific research workflows, revealing a critical gap in their ability to interact with specialized software and produce verifiable results.
Over 70% of sparse autoencoder features exhibit fundamentally mismatched input and output semantics, breaking the common assumption that what activates a feature directly mirrors its downstream causal effect.
CogEvol slashes learning artifact generation time from minutes to seconds while ensuring high reliability and quality.
DeepWeaver transforms the way LLMs synthesize evidence, leading to answers that are not only more comprehensive but also better grounded in citations.
Directly verbalizing sparse autoencoder features from LLM representations transforms how we interpret model behavior, making explanations more efficient and insightful.
SFT leads to task conflicts that can cripple multi-task learning, while RL's variance-limited updates enable seamless task coexistence.
