Homeopathic Learning: Diluting poisoned fine-tuning data can strengthen its effect
Matt Clarke, @jan_dubinski_, @jameschua_sg, @SimonSchrodi, @OwainEvans_UK, Alex Cloud (@cloud_kx)
@AnthropicAI
Double Blind Evals: Resolving the Dual Confidentiality Dilemma in AI Safety Auditing
@iamtrask, @SolMessing, @vaipahwa, @pegahbyte, Rene Kolga, Austin Frantz, Andrew Tash, Kate Thomas, Sean McGregor, @GeorgeBalston, @prpaskov, @Miles_Brundage, @m_c_elish, @KLdivergence, et al.
Towards Understanding Score-Seeking in LLMs
Aghyad Deeb (@aghyadd98), @RyanGreenblatt, @alextmallen
@redwood_ai
Mass Producing Training Data with Surprising Side Effects
@yaowenye123, @JacobSteinhardt, Sarah Schwettmann (@cogconfluence), @cassidy_laidlaw
@TransluceAI
Poisoning Attacks on LLMs Require a Near-constant Number of Poison Samples
@AlexandraSouly, @javirandor, Ed Chapman, @alxndrdavies, Burak Hasircioglu, Ezzeldin Shereen, @CarlosMougan, @mavroudisv, @ErikJones313, @hkscy, Nicholas Carlini, @yaringal, @_robertkirk
@AISecurityInst
1
23
1,478
AI Evaluation Should Work With Humans
@jankulveit, Gavin Leech (@gleech), Tomáš Gavenčiak, @raymondadouglas
From AGI to ASI
Tim Genewein, @FranklinMatija, @AlexLerchner, @LaurentOrseau, @SamuelAlbanie, Adam Bales, Cole Wyeth, Stephanie Chan (@scychan_brains), @IasonGabriel, Joel Z. Leibo (@jzl86), @AllanDafoe, @mhutter42, Thore Graepel (@ThoreG), @ShaneLegg
@GoogleDeepMind
Interpreting Language Model Parameters
@BushnaqLucius, @danbraunai, Oliver Clive-Griffin, @BartBussmann, Nathan Hu, Michael Ivanitskiy, Linda Linsefors, @leedsharkey
@GoodfireAI
Debate Training Reduces Reward Hacking in RLAIF
Zachary Kenton (@ZacKenton1), Lili Janzer, @rorygreig1, Tian Huey Teh, Kirill Tyshchuk, Jonah Brown-Cohen, @HarriLEdwards, @sen_r, @noahysiegel, @natashajaques, @rohinmshah
@GoogleDeepMind










