Greedy decoding
PulseAugur coverage of Greedy decoding — every cluster mentioning Greedy decoding across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
LLMs hallucinate software packages, posing supply chain risks; defenses evaluated
Researchers are investigating the problem of Large Language Models (LLMs) hallucinating non-existent software packages when generating code. This issue poses a risk to the software supply chain by creating exploitable v…
-
New attack method bypasses AI jailbreak defenses by combining code and search techniques
Researchers have demonstrated that a combination of a code-completion encoding and a best-of-N search can effectively bypass self-check jailbreak defenses in AI models, achieving success rates as high as 67% on some tar…
-
New decoding method preserves LLM refusal behavior under high-temperature sampling
Researchers have developed a new decoding method called Refusal-Gated Decoding to maintain an LLM's safety guardrails when using high-temperature sampling. This technique aims to increase output diversity without sacrif…
-
LLM grammar correction improved with edit-level majority voting
Researchers have developed a new method to address the over-correction problem in large language models used for grammatical error correction. Their training-free inference technique involves generating multiple correct…