The Archive

I've fine-tuned Qwen 3.5 0.8B on the dataset provided by Pangram with their EditLens paper. It's available via a [Chrome extension](https://chromewebstore.google.com/detail/slop-hammer/gfjdmhfokmhedlgfggmmgchpppmhkdgg); you can just click selected text and it's going to give you the probability distribution of how likely it is AI-generated. It takes under 1s on my M1 MacBook Pro. Pangram did release Llama 3.2 3B trained on their dataset, but I found this model slightly too legacy (too big for the capabilities). Qwen 0.8B (base) ended up being as good after roughly 20h of fine-tuning on a sin...

Channel-wise Vector Quantization replaces patch-wise image tokenization with channel-wise quantization for autoregressive visual models.

DiscoverPhysics benchmark tests LLM reasoning by having agents discover physics laws in simulated worlds with non-standard dynamics.

Claw-Anything benchmark evaluates LLM agents as always-on assistants with access to long-horizon histories and interdependent backend services.

VeriTrace proposes explicit feedback mechanisms for research agents to evolve mental models and prevent error propagation in uncertain information.

Auto Benchmark Audit uses agentic framework to systematically audit AI benchmarks for hidden dependencies, specification gaps, and grading flaws.

Wasserstein Policy Gradient for entropy-regularized RL proves global convergence using optimal-transport geometry for continuous control.

StakeBench evaluates financial language understanding using real market commitments from 560K+ Polymarket and Manifold comments instead of human labels.

Active Query Synthesis for preference learning uses confidence-aware response model to reduce labeling cost for pairwise comparisons.

WhoSaidIt applies human-LLM collaborative re-annotation to stabilize multilingual speaker-attribute labels via disagreement-focused sampling.

WSADBench unifies weakly supervised anomaly detection evaluation across incomplete, inexact, inaccurate supervision for 36 algorithms and 4 modalities.

Developer shares prompt-based skills to make Claude challenge product ideas rather than automatically validate them via custom instructions.

Conditional kernel ridge regression with unpenalized features via conditionally positive definite kernels; primarily theoretical ML contribution.

Paris 2.0: first decentralized video generation model trained without GPU clusters, extending prior Paris 1.0 image work.

llama.cpp adds fast Walsh-Hadamard transform (FWHT) for CUDA, yielding 1–2% prompt-processing and 7–9% token-generation speedups with quantized KV-cache.

Neuronal Stochastic Attention Circuit for uncertainty in continuous-time representation learning using C.elegans-inspired circuits.

Neural operator surrogates for Bayesian inverse design in CFD to accelerate MCMC sampling for aerodynamic geometry inference.

Retrying vs resampling in AI control: retrying leaks exploitable monitor rationale; resampling preserves safety without information leakage.

Chris Olah comments on papal encyclical; off-topic for AI industry professionals unless directly addressing AI ethics.

Multi-objective textual gradient optimization for LLM judges: gradient conflicts cause failure modes; six decomposition modes tested.

Uncertainty quantification for activation oracles interpreting LLM internals; bootstrap frequency is best-calibrated confidence method.

L2IR: graph fraud detection combining GNNs and LLMs to infer fraudster intent behind suspicious connections.

DRScaffold: training lightweight vision-language models for dense-scene reasoning with explicit grounding between inference steps.

Transformer paper co-author argues for moving beyond transformer architecture; Pathway's post-transformer research gaining attention from technical community.

Peak-then-collapse failure in GRPO tool-use training on knowledge graphs: four recurring failure modes shift rather than resolve.

CityRep benchmark evaluates urban representation learning across cities and modalities using spatially-structured splits to prevent data leakage.

MLP-LDRU architecture improves length generalization in neural networks via log-depth recurrent units with associativity-biased operators.

SKILD unifies image generation and super-resolution via scale-invariant diffusion in K-space, leveraging scale invariance in natural and physical systems.

CausaLab environment benchmarks LLM agents on interactive causal discovery with validation of both solutions and underlying causal mechanisms.

Boston Dynamics and Hyundai plan to train Atlas humanoid robot using football video learning, releasing progress via 'School of Football' documentary series.