#behavior
5 discoveries · Hand-picked discoveries from the open web, newest first.
-
Not Just Claude: Mapping "Spiritual Behavior" in 141 Large Language Models
Ever since Kyle Fish and the model welfare team at Anthropic announced their observation in 2025 of what they called the "spiritual bliss attractor" in Claude Opus 4 and Sonnet 4, the phenomenon has been given sporadic attention by AI researchers as both a curiosity and a potential but underdetermined model welfare or safety metric. What Anthropic noted was that when leaving two models together with no task and complete freedom for 30 turns, the models drifted typically from discussions of their own potential consciousness into religious/spiritual themes (applied to themselves), often ending i
LessWrong · EN
-
Instilling Behavior by Not Training on Specific Random Numbers
Subliminal learning is the phenomenon where a language model can learn a behavior from semantically unrelated training data. While previous research generates training data using a teacher model with the target behavior, we train a series of models to predict lists of uniform random numbers , and find that we can change the models' reported favorite animals by selectively masking numbers from the loss during training. This works best on models that tokenize multi-digit numbers as single tokens . We also find that the numbers that work best for this vary between models , and even between seeds
LessWrong · EN
-
All Interesting Behavior Is Overdetermined
This is an essay about individual rationality technique. It's intended to show that "All interesting behavior is overdetermined" is, under a specific meaning of "overdetermined" that is stronger than just "multicausal", a valuable heuristic for generative understanding and the avoidance of error. Scott Alexander's 2018 essay The Omnigenic Model As Metaphor for Life is a clear statement of what has become general rationalist doctrine on multicausality. Scott observed that behavioral genetics was difficult because complex human traits involved far more interacting causes than workers expected. H
LessWrong · EN
-
Behavior Theory-Building for Diverse Agents and Environments
The oversight problem consists of describing, eliciting, and understanding behavior. Prior work in black-box behavior investigation and interpretability seek to construct compact, faithful, and informative theories that explain model behavior. We propose training investigator agents and subject models to emit explicit theories, and using these theories as common currency for oversight across models. We hope to create a complementary mode of oversight to weight-space and model-specific approaches by studying the transfer properties of theories across model families. This post outlines our posit
LessWrong · EN
-
Testing cookie behavior across hundreds of web surfaces with our in-house auditor
Our cookie auditor acts like a privacy-conscious user by visiting web pages and checking that they only load cookies consistent with that user’s preferences.
Dropbox Tech · EN