Reproducing "Natural Emergent Misalignment from Reward Hacking" (MacDiarmid et al., Anthropic 2025) with open-source models. Includes reward-hackable RL environments, misalignment evaluations, training configs, and evaluation scripts. Models trained on OLMo (7B, 32B) and GPT-OSS (20B, 120B).
☆36Oct 5, 2026Updated last month
Alternatives and similar repositories for reward-hacking-misalignment
Users that are interested in reward-hacking-misalignment are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- ☆36Updated this week
- ☆47Feb 18, 2026Updated 8 months ago
- Zero-dependency local web app for browsing Claude Code session transcripts☆15Updated this week
- A public-domain dataset of prompts and scenarios for evaluating compliance with the OpenAI Model Spec.☆17Mar 24, 2026Updated 7 months ago
- Code repo for the model organisms and convergent directions of EM papers.☆90Sep 22, 2025Updated last year
- Managed Database hosting by DigitalOcean • AdPostgreSQL, MySQL, MongoDB, Kafka, Valkey, and OpenSearch available. Automatically scale up storage and focus on building your apps.
- A toolkit that provides a range of model diffing techniques including a UI to visualize them interactively.☆88Oct 25, 2026Updated 2 weeks ago
- James' cookbook of evaluations and finetuning experiments☆36Feb 19, 2026Updated 8 months ago
- Extract residual-stream activations and apply steering vectors (including activation oracles) to any vLLM model during inference.☆142Oct 13, 2026Updated last month
- Official Inspect Implementation for "ImpossibleBench: Measuring LLMs' Propensity of Exploiting Test Cases"☆81Oct 13, 2026Updated last month
- ☆67Jul 4, 2025Updated last year
- ☆27Jul 8, 2026Updated 4 months ago
- ☆61Feb 11, 2025Updated last year
- ☆14Jun 10, 2025Updated last year
- A toolkit for embedding text datasets with sparse autoencoders☆32Mar 24, 2026Updated 7 months ago
- Managed Kubernetes at scale on DigitalOcean • AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- R package for automated analysis of serological data☆10Sep 26, 2026Updated last month
- ☆20Aug 17, 2026Updated 2 months ago
- Pytorch implementation on OpenAI's Procgen ppo-baseline, built from scratch.☆14May 17, 2024Updated 2 years ago
- ☆17May 21, 2026Updated 5 months ago
- Open Character Training☆107Apr 4, 2026Updated 7 months ago
- cheatbench.ai☆20Updated this week
- Fast, standardized, and easy-to-use interpretability engine.☆45Oct 27, 2026Updated 2 weeks ago
- [ICLR 2025] On Evluating the Durability of Safegurads for Open-Weight LLMs☆14Jun 20, 2025Updated last year
- ☆16Jan 2, 2026Updated 10 months ago
- Wordpress hosting with auto-scaling - Free Trial Offer • AdFully Managed hosting for WordPress and WooCommerce businesses that need reliable, auto-scalable performance. Cloudways SafeUpdates now available.
- A python library built on top of Inspect AI to support Control research and evaluations, by Redwood and EquiStamp☆46Updated this week
- ☆15Nov 3, 2025Updated last year
- Monorepo that hosts and serves the iliad intensive content☆18Updated this week
- MishformerLens intends to be a drop-in replacement for TransformerLens that AST patches HuggingFace Transformers rather than implementing…☆10Oct 7, 2024Updated 2 years ago
- ☆20Jan 21, 2023Updated 3 years ago
- ☆25May 25, 2024Updated 2 years ago
- ☆20Jun 3, 2026Updated 5 months ago
- Training LLMs to Report Their Learned Behaviors☆31Apr 28, 2026Updated 6 months ago
- ☆344Jan 12, 2026Updated 10 months ago
- Managed hosting for WordPress and PHP on Cloudways • AdManaged hosting for WordPress, Magento, Laravel, or PHP apps, on multiple cloud providers. Deploy in minutes on Cloudways by DigitalOcean.
- Better youtube recommendations☆25Feb 19, 2026Updated 8 months ago
- TikZ version of the standard model graphic from particle fever☆11Apr 23, 2020Updated 6 years ago
- ppx_system is a syntax extension to known operating system at compile time☆12May 9, 2023Updated 3 years ago
- A python sdk for LLM finetuning and inference on runpod infrastructure☆31Updated this week
- A fork of openai/miniF2F adapted to Lean 4, with corrections to formalizations and informal descriptions. for human readers.☆15Apr 23, 2026Updated 6 months ago
- A suite of interpretability tasks to evaluate agents using Scribe for notebook access☆18Oct 2, 2025Updated last year
- LLM Uncertainty Quantification with Attention Chain☆11Mar 24, 2025Updated last year
Are these results useful?