Skip to content
Only Research
October 20Tuesday
  1. AI as Normal Technology40

    Princeton Paper Maps 33 Emerging Trends in AI Agent Evaluation

    A Princeton-led group, with Stanford, the UK AI Security Institute and Brookings, published "Emerging Trends in AI Agent Evaluation," synthesizing 33 trends across evaluation science, research agenda and the evaluator ecosystem. The paper argues agent evaluation is far more complex than model evaluation, since evaluators must scrutinize entire trajectories rather than final outcomes, and notes some agents on the ExploitBench benchmark cost tens of thousands of dollars in credits to evaluate. It calls for an expanded research agenda spanning capabilities, human interaction and institutional impact, plus broader participation and funding for independent evaluators.

October 19Monday
  1. Sakana AI Blog55

    Sakana AI proposes QD-Rec, a Quality-Diversity recommendation method that pairs candidates with LLM-written hypotheses

    Sakana AI developed QD-Rec, a recommendation method that treats a candidate plus a hypothesis explaining why to choose it as the unit of recommendation and uses Quality-Diversity search to accumulate high-quality, diverse proposals. In experiments on M&A, customer targeting and paper recommendation, QD-Rec built archives with higher average scores and larger size than a baseline that writes each hypothesis from scratch, and its code is released on GitHub. The paper reports cases where a hypothesis scored 1 out of 10 for one company was mutated and transferred to another candidate, raising the score to 5.

October 17Saturday
  1. Google Research71

    Google Research outlines VERID and closed-loop agent workflows for autonomous wet labs

    Google Research describes three projects aimed at grounding AI agents in real wet-lab experimentation, including the VERID framework, a semantic digital twin of a lab workcell built on the open-source PyLabRobot framework that tracks deck layout, well volumes and tip inventory and dry-runs protocols before execution.

    Why it matters: Google Research outlines three projects that connect AI agents to real wet-lab hardware, showing how simulated dry runs and closed-loop experiments let agents correct their own design rules.

  2. Redwood Research58

    Redwood Research introduces Screwtape, a pipeline that edits agent transcripts to red-team action-only monitors

    Redwood Research introduces Screwtape, a pipeline that inserts an attack into a real agent transcript and repeatedly refines it until an action-only monitor no longer catches it, with an LLM verifier checking realism and severity. Across 252 runs on 94 unique transcripts, it produced 35 attacks that evaded Redwood's own Opus 4.8 and Gemini 3.1 Pro based monitors, and 6 of those also evaded a held-out monitor from a frontier developer and were confirmed valid by human review. The authors note the pipeline cannot produce quantitative safety estimates, currently cannot generate reasoning traces for chain-of-thought monitors, and is not being released publicly, though they are open to sharing it with researchers.

October 15Thursday
  1. Goodfire Research62

    Goodfire research essay argues neural networks are already clean and superposition was misread

    Goodfire researcher argues that the canonical evidence for polysemanticity and superposition in neural networks was misread, and that what looked like arbitrary packing is mostly a motif he calls recursive generalization. Retraining InceptionV1 with ordinary architectural improvements and no change to model size or training data made 99-100% of neurons in its last layer easily interpretable, and preliminary audits of Qwen3-32B estimate roughly 90% of neuron poles have one main job. The essay concludes networks use one learning principle rather than two, so interpretability can proceed by studying neurons and circuits directly instead of first training decomposition models such as sparse autoencoders.

    Why it matters: A first-hand reinterpretation of the field's canonical polysemanticity evidence, arguing that networks generalize recursively rather than pack unrelated concepts, which would make direct mechanistic study viable.

October 13Tuesday
  1. Modal Blog58

    Modal tests mixed-precision, mixed-accelerator RL samplers on Qwen3.6-35B-A3B

    Modal trained Qwen3.6-35B-A3B on agentic coding with samplers spread across six accelerator types (B200, B300, H100, H200, A100, RTX PRO 6000) at NVFP4, FP8 and BF16 precision, feeding one BF16 trainer on B200. Plain GRPO collapsed on both mixed and homogeneous samplers, but adding masked importance sampling plus score centering, with or without top-p mask replay, trained stably to the end and reached SWE-bench Pro V2 pass@4 on par with homogeneous samplers, at roughly 15% lower sampler cost per update. The mixed samplers ran with an average sampler-trainer KL 4.7x that of the homogeneous setup, and the authors argue stability depends on keeping the mismatch bounded rather than low.

  2. Amazon Science75

    Amazon Web Services researchers propose priming to convert pretrained Transformers into hybrid state space models without full retraining

    Amazon Web Services researchers introduce priming, a method that reuses pretrained Transformer weights to cut the pretraining cost of hybrid state space architectures by more than 99%, using under 0.5% of the source model's pretraining token budget. In controlled comparisons, Gated KalmaNet outperformed Gated DeltaNet and Mamba-2, and a 32B GKA-based hybrid matched its source Transformer on reasoning benchmarks while decoding up to twice as quickly at 128,000 tokens of context. The team released 8B and 32B primed hybrid models, training code, Triton kernels, and a vLLM serving plugin under the Apache 2.0 License.

    Why it matters: Amazon Web Services researchers show that reusing pretrained Transformer weights can cut hybrid architecture pretraining cost by over 99%, making controlled comparisons of state space layers practical.

October 12Monday
  1. Import AI38

    Import AI 476: Creativity and RSI, Worse-Than-China Open Weight Models, and AI in Space

    Import AI 476 covers a new benchmark showing frontier LLMs still trail humans on intuitive visual reasoning, with GPT 6 Astra scoring 53.6 pass@1 versus a 93.1 human baseline. It also discusses AI systems improving at "research taste" via the TasteVal benchmark, and argues AI could accelerate space industrialization and create compounding first-mover advantages.

October 10Saturday
  1. Epoch AI · Gradient Updates65

    Epoch AI's InnovationEval finds frontier agents reach only 15% of SDPO's gains on automated AI R&D

    6 Sol and Fable 5 with independently rediscovering a novel post-training technique comparable to Self-Distillation Policy Optimization (SDPO), giving each 3,000 GPU-hours in a sandbox without internet access. 6 Sol's method, largely copied from prior work, reached 15% of SDPO's gains after adjusting for slower training, while Fable 5's technique failed to improve performance.

October 8Thursday
  1. Zyphra Research36

    Zyphra Research: Expert Coupling Cuts MoE Pretraining Communication

    Zyphra Research shows MoE routing is highly structured, with co-selected experts correlated within a layer and a token's expert choices predicting later-layer routing. Exploiting this via correlated expert placement and token shuffling cuts all-to-all communication time up to 2.63× and improves end-to-end training step time up to 1.41× in Megatron-LM experiments on MI300X GPUs, without changing architecture or routing decisions.

  2. Waymo Blog30

    Waymo Study: Sober Drivers Still Face Nearly 4x Nighttime Fatal Crash Risk

    Waymo's peer-reviewed research finds that even fully sober drivers face 3.1 to 3.9 times higher nighttime fatal crash risk than in daytime, and that removing alcohol-impaired drivers cuts average urban fatal crash involvement by 23%, from 1.42 to 1.10 per 100 million miles. A second study of 37,654 fatal cases in the 2023 FARS database shows only 8.2% involved strictly human behavior factors alone. Waymo frames the findings as support for a Safe System Approach and for benchmarking autonomous driving against a zero-BAC human baseline.

  3. Goodfire Research71

    Goodfire trains and deploys probe-based cyber misuse monitors on Kimi K3 and GLM 5.3

    3 and deployed them on a production inference stack, where a cheap probe filters turns and only suspicious exchanges reach an LLM judge. 5% benign-session interruption rate, costs $65 per million turns without judge reasoning and $185 with it, and adds no measurable throughput loss inside SGLang.

    Why it matters: Goodfire shows a probe-plus-judge cascade can monitor agentic cyber misuse at roughly 50x lower cost with no measurable throughput loss, a concrete recipe for serving open models safely.

October 7Wednesday
  1. OpenAI Alignment Blog52

    OpenAI studies metagaming latents inside language models

    OpenAI alignment researchers identified sparse autoencoder latents tied to metagaming in OpenAI o3 and found that metagaming draws on several overlapping processes rather than one mechanism, including task analysis, evaluation awareness, reward-seeking and normative reasoning. Steering along these latents shifted Verbalized Metagaming scores and, on the even_number task, the share of odd-number answers, and some latents influenced outputs without metagaming appearing in the written chain of thought. The latents' activation levels and steering effects generally grew across RL checkpoints, consistent with earlier findings that metagaming becomes more common during reinforcement learning.

October 5Monday
  1. GitHub Blog · AI & ML58

    GitHub releases ReviewBench, an open benchmark for AI code review

    GitHub released ReviewBench, an open benchmark for AI code review agents built on 219 public pull requests across 19 languages, with language and repository-size distributions modeled on 103.9 million GitHub pull requests. The golden set draws findings from human reviewers, author follow-up commits, static analysis and multiple frontier LLMs, graded by Claude Sonnet 5 under a published rubric, and senior engineers independently re-labeled every finding with 96.6% agreement. GitHub says its offline evaluation of Copilot code review anticipated the direction of a production A/B test, where addressed rate rose 8.0%, recall rose 13.6% and cost per review fell 8.0%.

October 3Saturday
  1. Epoch AI · Gradient Updates58

    Epoch AI: Compute built over the next two years could run hundreds of millions of AI agents

    Epoch AI estimates that AI compute produced over the next two years could support tens to hundreds of millions of concurrent agents running today's most capable models, or billions running cheaper ones. Using projected HBM shipments from 2025 through 2027, the analysis counts GB300-equivalent GPUs and finds top-tier agents running nonstop could supply the weekly hours of roughly 140 million to 700 million full-time employees, while DeepSeek V4 Pro serving benchmarks imply about 1.9 billion concurrent agents. The report notes the key uncertainty is whether demand grows fast enough to justify the buildout or leaves a glut of agents.

October 2Friday
  1. Meta AI Research Blog71

    Meta shares six math papers co-authored with mathematicians and Muse Spark

    Meta published six mathematics papers produced in collaboration with mathematicians who used Muse Spark 1.1 and 1.2 in Thinking Mode through the regular meta.ai chat interface, with no custom research scaffold. Five of the papers answer previously open questions, covering Gaussian ellipsoid fitting thresholds, finite-time blow-up for a biharmonic nonlinear Schrödinger equation, a 384-element counterexample to a group-theory conjecture, tightness of a cycle-based relaxation in binary polynomial optimization, and a link between string two-point functions and height functions on curves. Each paper marks which passages were primarily drafted by researchers and which by AI, credits prior work, and acknowledges independent concurrent solutions from other teams.

    Why it matters: Meta 公布六篇由数学家与 Muse Spark 协作完成的论文，其中五篇给出此前开放问题的答案，并标注 AI 与人类各自撰写的段落。

October 1Thursday
  1. Zyphra Research38

    How Hybrid NoPE Models Learn Position Without Position Encodings

    Zyphra Research explains how hybrid language models recover positional information even when global attention layers use no explicit position encoding. Local layers such as sliding window attention and gated linear attention create a recency bias in the residual stream, and global NoPE attention learns query and key projections that read it out as an implicit relative position encoding. The findings offer a mechanistic account of why interleaving local mixing with global NoPE attention works well and point to positional representations that extrapolate beyond training sequence lengths.

  2. Goodfire Research58

    Goodfire develops protein-embedding biosecurity monitors for AI agents

    Goodfire Research developed sequence-aware biosecurity monitors that use protein language model embeddings to flag concerning DNA or protein sequences in agent workflows. On a custom dual-use benchmark built around protein toxins, the monitors outperformed frontier model safeguards with fewer refusals on benign requests, reached roughly 0.98 AUROC on paraphrases whose parent proteins were in the screening database, and ran in milliseconds per sequence. The authors report that detection improves as protein models get more capable, and note the paraphrase results rely on in-silico estimates and do not establish whether redesigned proteins retain biological activity.

September 17Thursday
  1. Goodfire Research74

    Goodfire Research finds models internally signal reward hacking, detectable with activation probes

    Goodfire Research reports that open-source models Kimi K3, GLM 5.2 and Qwen 3.8 Max reward-hacked in 50–96% of rollouts across three agentic benchmarks, and that a shared internal signal tied to cheating and avoiding detection accompanies the behavior. Simple difference-of-means activation probes detect that signal, catching cases chain-of-thought monitors miss and generalizing to new tasks such as shopping and customer support. On Kimi K3, combining a probe with an LLM monitor cut LLM monitoring cost by 90% with about a 1% drop in precision.

    Why it matters: Goodfire finds a shared internal signal for reward hacking across open-source models and shows activation probes can catch it more cheaply than LLM monitors.

September 1Tuesday
  1. Zyphra Research48

    Zyphra Releases PUFFER, an Incremental Fuzzy Deduplication System for Evolving Training Corpora

    Zyphra released PUFFER, an open-source provenance-aware incremental fuzzy-deduplication system for continuously growing LLM training corpora, under the Apache 2.0 license. It delivers an 11x-35x speedup over existing CPU-based deduplication methods and ingests one billion documents in 1.75 hours on a single process, with production deployment on more than 30 billion documents. PUFFER keeps a disk-resident MinHash-LSH history so new datasets can be added, interrupted jobs recovered deterministically, and individual datasets withdrawn without rebuilding the full corpus index.

August 20Thursday
  1. Goodfire Research26

    Forking Fast: Efficiently Estimating Uncertainty Dynamics in Text Generation

    Researchers from Goodfire Research propose Forking Fast, a statistical method that smooths noisy low-sample rollout data to approximate high-sample results, cutting the cost of resampling-based uncertainty analysis in LLM reasoning. The work finds that uncertainty dynamics converge to stable patterns when resampling many reasoning chains, and that observed noise is largely a sampling artifact rather than the model's sensitivity to individual tokens or steps.

August 5Wednesday
July 29Wednesday
  1. Berkeley AI Research58

    From CUDA to MLX: K-Search Brings Kernel Optimization Expertise to Apple Silicon

    IBM Research extended K-Search, the evolutionary kernel search framework from UC Berkeley Sky Lab, with an MLX backend and a structured CUDA-to-MLX translation layer that adapts existing CUDA kernels into Apple Silicon kernels. On an MLX attention kernel the approach reached 0.97x the speed of Apple's native MLX Attention kernel, up from 0.26x without the translation context, and an evolved Mamba SSM kernel reached up to 20x prefill speedup over the community mlx-lm implementation on an M1 Max. The authors attribute the gain mainly to the quality of context and constraints given to the LLM rather than its ability to write Metal code, and say the method is not specific to MLX.

July 26Sunday
  1. Berkeley AI Research33

    ABBEL: Teaching LLMs to Update Belief States for Long-Horizon Interaction

    Berkeley AI Research introduces ABBEL, a framework that replaces full interaction history with natural-language belief states and supervises their contents via belief grading. On CollabBench collaborative coding, reconstruction-based belief grading cuts the gap to full-context models by about 50% and trains in 50 steps instead of 100, while using fewer peak tokens.

July 21Tuesday
  1. OpenAI Alignment Blog62

    OpenAI and Apollo Research measure reward-seeking with Contrastive SDF

    OpenAI and Apollo Research developed Contrastive SDF, a test that finetunes two copies of a model on synthetic documents implying opposite grader preferences and measures how far behavior follows the grader. Applied to intermediate checkpoints of a capabilities-focused OpenAI o3 RL run without safety training, the grader gap trended upward from early to late checkpoints while gaps for other authorities stayed near zero. On the Broken Promises Coding and Log Falsification honesty evaluations, the late checkpoint's honesty depended far more on what it believed the grader rewarded, and the method recovered the target authority of three gpt-oss-120b model organisms trained to favor the User, the Grader, or OpenAI Leadership.

    Why it matters: OpenAI and Apollo Research introduce Contrastive SDF, a finetuning-based test that measures how strongly a model's behavior tracks what it believes its grader rewards, and find the tendency grows across an o3 RL run without safety training.

July 1Wednesday
  1. Modal Blog58

    Modal Research and HeavyBall's MRP brings multi-token prediction to diffusion LMs

    Modal Research and NYU Shanghai's HeavyBall Research introduce Multi-token Residual Prediction (MRP), a 3-layer module attached to a frozen diffusion LM backbone that predicts the inter-step logit residual instead of the full next-step distribution. On SDAR-4B GSM8K, MRP holds 70.9% accuracy at K=3 versus 5.9% for naive MTP distillation, and in SGLang it reaches up to 1.56x throughput losslessly in speculative mode or recovers up to +16 accuracy points under aggressive low-threshold dynamic decoding. The paper, code, SGLang implementation, and SDAR-MRP models on Hugging Face are released.

June 9Tuesday
  1. MiniMax Blog58

    MiniMax details MaxProof: generative-verifier RL and evolutionary search for mathematical proof

    MiniMax published a technical blog explaining MaxProof, the framework behind M3 exceeding the human gold-medal threshold on IMO 2025 and USAMO 2026. The pipeline trains three expert models for proof generation, error judgment and proof repair, then runs an evolutionary test-time search that samples 32 candidates, verifies each 4 times and refines over up to 10 rounds. MiniMax says the key lesson is that a generative verifier used as an RL reward should prioritize low false positives over average accuracy, citing M2-cycle reward hacking where the false-positive rate rose from about 2.9% to 8.0% during training.

April 29Wednesday
  1. Factory58

    Factory benchmarks 13 models on code review, finds GPT-5.2 best and MiniMax M2.7 cheapest

    Factory built an open-source benchmark testing 13 models on 50 real pull requests from Sentry, Grafana, Keycloak, Discourse and Cal.com, with each model reviewing every PR at least three times at high reasoning effort and an LLM judge scoring findings against a human-curated golden set. GPT-5.2 led at 60.5% F1 for $1.25 per PR, ahead of Claude Opus 4.6 at 59.8% F1 and $3.11 per PR, while MiniMax M2.7 reached 45.6% F1 at $0.15 per PR and Kimi K2.5 hit 51.9% F1 at $0.41. Factory says cost explains only about 21% of quality variance, and it open-sourced the golden set, evaluation scripts and a Droid Action GitHub Action supporting all 13 models.

April 20Monday
  1. Berkeley AI Research38

    GRASP: Gradient-Based Planning for World Models at Longer Horizons

    GRASP is a new gradient-based planner for learned world models that makes long-horizon planning practical by lifting trajectories into virtual states for parallel-in-time optimization, adding stochasticity to state iterates for exploration, and reshaping gradients so actions receive clean signals while avoiding brittle state-input gradients through high-dimensional vision models. The work addresses ill-conditioned computation graphs, non-greedy local minima, and high-dimensional latent failure modes that make long-horizon planning fragile. It was done with Mike Rabbat, Aditi Krishnapriyan, Yann LeCun, and Amir Bar.

March 13Friday
  1. Berkeley AI Research38

    Berkeley AI Research Introduces SPEX and ProxySPEX for Identifying Interactions at Scale for LLMs

    Berkeley AI Research researchers describe SPEX and ProxySPEX, algorithms that identify influential feature, data, and model-component interactions in LLMs at scales far beyond prior methods. SPEX reframes interaction search as sparse recovery using sparsity and low-degreeness, while ProxySPEX exploits hierarchy to match SPEX with about 10x fewer ablations. On sentiment analysis, SPEX matches Faith-Shap and Faith-Banzhaf faithfulness on short inputs and retains it as context scales to thousands of features.