Tessera
Folgen

Daniel Scalena

@danielsc4.it

Intern of TS @Cohere | PhDing @unimib 🇮🇹 & @GroNlp 🇳🇱, interpretability et similia danielsc4.it

24 Posts434 Follower185 Folgt
Daniel Scalena@danielsc4.it · last mo.

I'd never have guessed models commit to their final answer this early, often within the first 20% of reasoning, across math/logic tasks and model families. The rest is mostly hedging that doesn't change their mind. And turns out they encode this internally, we can decode it! 🧵👇

Sara Candussio@saracandussio.bsky.social · last mo.

Are all the CoT steps necessary? In our latest paper, we find evidence for the existence of a commitment boundary, marking a sharp transition from no/mid guesses to the model final answer across various reasoning tasks and model families. Thread 🧵👇

Daniel Scalena@danielsc4.it · 10mo ago

You can easily save up to 65% of compute while improving performance on reasoning tasks 🤯 👀 Meet EAGer: We show that monitoring token-level uncertainty lets LLMs allocate compute dynamically - spending MORE on hard problems, LESS on easy ones. 🧵👇

Bild
Daniel Scalena@danielsc4.it · last yr.

📢 New paper: Applied interpretability 🤝 MT personalization! We steer LLM generations to mimic human translator styles on literary novels in 7 languages. 📚 SAE steering can beat few-shot prompting, leading to better personalization while maintaining quality. 🧵1/

Bild