We introduce PiCSAR (Probabilistic Confidence Selection And Ranking)💡: A simple training-free method for scoring samples based on probabilistic confidence, selecting a reasoning chain with the highest confidence from multiple sampled responses. ✏️PiCSAR is generalisable across

Picking reasoning chains by confidence works!

🚀 Excited to see our work on PiCSAR out! Thrilled to have Joshua as a co-author — and even more thrilled that he’ll be joining my group this academic year. Big things ahead!

Happy to share that our work is accepted by NeurIPS 2025 LAW workshop: Bridging Language, Agent, and World Models! ☀️ May see you in San Diego.

NO verifiers. NO Tools. Qwen3-4B-Instruct can match DeepSeek-R1 and o3-mini (high) with ONLY test-time scaling. Presenting Recursive Self-Aggregation (RSA) — the strongest test-time scaling method I know of! Then we use aggregation-aware RL to push further!! 📈📈 🧵below!

Expected Attention compresses the KV Cache by estimating the attention score from future queries! Both during prefilling & decoding 🚀 All code released in our KVPress library & Leaderboard!

Thrilled to share our latest research on verifying CoT reasonings, completed during my recent internship at FAIR AI at Meta. In this work, we introduce Circuit-based Reasoning Verification (CRV), a new white-box method to analyse and verify how LLMs reason, step-by-step.

Our Imperial NLP community is ready for #EMNLP2025 ✈️ Check out the papers we’ll be presenting next week in Suzhou and come say hi!

🦄 Googling how to build an additional MLP head somehow leads to tutorials on rerooting My Little Pony figures🦄