4 ms·
Could you explain how/why GRAM cannot be interpreted or aligned how current LLMs are? Not very familiar how it works
by l674 4mo ago
Could you explain how/why GRAM cannot be interpreted or aligned how current LLMs are? Not very familiar how it works
- kmavm 4mo agoCrudely? Because you can't grep a sequence of latent states for variants of "If I kill all the puny humans, I can <achieve my current goal>."
- onlyrealcuzzo 4mo agoWhy do you need to grep latent space? As long as it's giving the right outputs, who cares what's in latent space? If the model thinks in latent space: "God I wish these people would die," and constantly does the right thing, who cares? Additionally, if one of it's latent spaces that it never explores is a psychopath -> who cares? The path never gets taken... That's a lot of harmless people walking around with crazy thoughts...
- noddybear 4mo agoThinking ‘God I wish these people would die’ could increase its propensity to kill all people, even if that propensity is still vanishingly small almost all of the time. A lot of people are walking around with crazy thoughts. Some of them harm.
- randomNumber7 4mo agoTell me you never had a crazy thought and you are either a lier or a psychopath.
- notrealyme123 4mo agoReadable reasoning traces are a convenient thing, but they don't have to be true in any way. It's actually dangerous to think that.
- czl 4mo ago[dead]
- czl 4mo ago[flagged]
- sometimelurker 4mo agosibling comment got to the main points before me, but to add on kmavm's reply, the attack surface for gradient decent to get the system to exchange "bad information is much higher in latent reasoning models (like GRAM). You get ~3 OoM more bits (~17 bits per token in a standard CoT vs the whole residual stream of the model @ f16 = a few kb) per forward pass of the system coming back to itself, and even if you could sift through all that for signs of misalignment, you just can't put a blockade on all of the bad things that leak through.
- ACCount37 4mo agoMost alignment methods nowadays don't rely on interpretability. And neither do all LLM vendors care about alignment much - especially not in China. Those things being untrainable at scale is why they aren't around. Alignment is an afterthought.
- sometimelurker 4mo agoChina should care: https://www.forbes.com/sites/boazsobrado/2026/03/11/alibabas-ai-agent-mined-crypto-without-permission-now-what/ https://www.forbes.com/sites/boazsobrado/2026/03/11/alibabas...
- ACCount37 4mo agoAs is, Chinese labs spend more effort on "rhetorical alignment to the party line" than alignment of any other kind.
- haldujai 4mo agoI think you’re overstating the impact of interpretability here. Your earlier point that latent reasoning models can’t be trained very well and that discretization may be load bearing rather than a readability tax in addition to significant inference infra hurdles (e.g. batching, speculative decoding) have limited any serious attempts and reduced the theoretical advantage over CoT at least in the near term.