Roland Pihlakas open source projects
Pinned Loading
Repositories
- Critical-Thinking-Annotator Public Forked from biological-alignment-benchmarks/Manipulative-Expression-Recognition
Fallacy and cognitive bias detector identifies and highlights reasoning fallacies and cognitive biases in text from human conversations and AI-generated responses. It promotes rational reasoning and clarity of mind. It mitigates harmful judgement and spreading of misinformation by detecting fallacies and biases in communication and thoughts.
- Agreement-and-Disagreement-Recognition Public Forked from biological-alignment-benchmarks/Manipulative-Expression-Recognition
ADR is a software that identifies and highlights agreements and disagreements in discussion forum messages as a response to the original post.
- Universal-Values-Assistant-Simulation Public
Universal Values Assistant Simulation. This code was developed based on research and ideas of Chad https://www.linkedin.com/in/chad-burghardt-723416142/ and coded by Roland https://github.com/levitation
- milgram-for-llms Public Forked from biological-alignment-benchmarks/milgram-for-llms
Four main takeaways: (1) LLMs are subject to pressure, they comply despite expressing distress; (2) LLMs are vulnerable to gradual boundary/value violations; (3) when LLMs refuse, they may ignore the response format requirements, so the query is retried; (4) we hypothesise there is a token pattern continuation attractor that might cause obedience.
- zoo_to_gym_multiagent_adapter Public Forked from biological-alignment-benchmarks/zoo_to_gym_multiagent_adapter
Enables you to convert a PettingZoo environment to a Gym environment while supporting multiple agents (MARL). Gym's default setup doesn't easily support multi-agent environments, but this wrapper resolves that by running each agent in its own process and sharing the environment across those processes.
- biological-alignment-gridagents-benchmarks Public Forked from biological-alignment-benchmarks/biological-alignment-gridagents-benchmarks
Safety challenges for AI agents' ability to learn and act in desired ways in relation to biologically and economically relevant aspects. The benchmarks are implemented in a gridworld-based environment. The environments are relatively simple, just as much complexity is added as is necessary to illustrate the relevant safety and performance aspects.
- bioblue Public Forked from biological-alignment-benchmarks/bioblue
Systematic runaway-optimiser-like LLM failure modes on Biologically and Economically aligned AI safety benchmarks for LLM-s with simplified navigation-free observation format. The benchmark themes include multi-objective homeostasis, (multi-objective) diminishing returns, complementary goods, sustainability.
- ai-safety-gridworlds-fork Public Forked from biological-alignment-benchmarks/ai-safety-gridworlds
Extended, multi-agent, and multi-objective (MaMoRL / MoMaRL) gridworld environments building framework based on DeepMind's AI Safety Gridworlds. This is a suite of reinforcement learning environments illustrating various safety properties of intelligent agents. It is made compatible with OpenAI's Gym/Gymnasium and Farama Foundation PettingZoo.
- DuplicateFileFinder Public
Duplicate file finder and de-duplicator. A tool that detects duplicate files and replaces them with symlinks to a shared file in a special shared files folder.
- MboxComparer Public
Memory efficient mbox-type mailbox files comparator for data integrity verification. It compares full content of emails (headers and body) after canonicalisation, and optionally tolerates differences in the lengths of blank-line runs. Optionally, it can handle mboxes with differing message orderings while still providing efficient resource usage.
Top languages
Loading…
Most used topics
Loading…