---
title: "AlphaSignal Post Training News"
url: "https://alphasignal.ai/topic/post-training"
description: "Latest AI and machine learning news about Post Training on AlphaSignal."
---

# AlphaSignal Post Training News

Latest AI and machine learning news about Post Training on AlphaSignal.

## Articles

- [Cohere's North Small Translate Beats DeepL Across 50 Languages With Downloadable Weights](https://alphasignal.ai/news/cohere-s-north-small-translate-beats-deepl-across-50-languages-with.md) (2026-10-03)
- [Ai2's AstaBrief 8B Writes Cited Research Reports 3.5x Faster in One Pass](https://alphasignal.ai/news/ai2-s-astabrief-8b-writes-cited-research-reports-3-5x-faster-in-one-pass.md) (2026-10-02)
- [Arena's FLUX.2-dev Beats Every Open-Source Image Model With 66% Win Rate](https://alphasignal.ai/news/arena-s-flux-2-dev-beats-every-open-source-image-model-with-66-win-rate.md) (2026-10-02)
- [Hugging Face and Liquid AI's LFM2.5 Jumps 12 Points Across Four Agent Harnesses](https://alphasignal.ai/news/hugging-face-and-liquid-ai-s-lfm2-5-jumps-12-points-across-four-agent-harnesses.md) (2026-10-02)
- [Cloudflare's Clef Beats Rivals With 38ms Typed Decisions for AI Agents](https://alphasignal.ai/news/cloudflare-s-clef-beats-rivals-with-38ms-typed-decisions-for-ai-agents.md) (2026-10-02)
- [Harvard's Projection Sampling Makes Standard Fine-Tuning Match Reinforcement Learning](https://alphasignal.ai/news/harvard-s-projection-sampling-makes-standard-fine-tuning-match-reinforcement.md) (2026-10-01)
- [MIT and Tel Aviv's LSL Stops AI from Forgetting Skills After Fine-Tuning](https://alphasignal.ai/news/mit-and-tel-aviv-s-lsl-stops-ai-from-forgetting-skills-after-fine-tuning.md) (2026-10-01)
- [Adaption Labs Turns a Single Prompt Into 20,000 AI Training Samples](https://alphasignal.ai/news/adaption-labs-turns-a-single-prompt-into-20-000-ai-training-samples.md) (2026-10-01)
- [Sharpening Tax Shows How RL Post-Training Quietly Kills AI Agent Retries](https://alphasignal.ai/news/sharpening-tax-shows-how-rl-post-training-quietly-kills-ai-agent-retries.md) (2026-10-01)
- [NVIDIA's PivotOPD Teaches AI Agents to Fix Early Mistakes Before They Spiral](https://alphasignal.ai/news/nvidia-s-pivotopd-teaches-ai-agents-to-fix-early-mistakes-before-they-spiral.md) (2026-09-30)
- [Kyutai's Structure-Agnostic Distillation Beats Dense Matching for Diffusion Image Models](https://alphasignal.ai/news/kyutai-s-structure-agnostic-distillation-beats-dense-matching-for-diffusion.md) (2026-09-30)
- [AC2 Beats GRPO on Math Proofs Using 2.5x Fewer Decoding FLOPs](https://alphasignal.ai/news/ac2-beats-grpo-on-math-proofs-using-2-5x-fewer-decoding-flops.md) (2026-09-30)
- [ETH Zurich's Adaptive Probes Beat DPO at Safety Without Breaking AI Transparency](https://alphasignal.ai/news/eth-zurich-s-adaptive-probes-beat-dpo-at-safety-without-breaking-ai-transparency.md) (2026-09-29)
- [University of Washington Finds LLM Agent Swarms Copy Well but Explore Poorly](https://alphasignal.ai/news/university-of-washington-finds-llm-agent-swarms-copy-well-but-explore-poorly.md) (2026-09-29)
- [Apple's RLTL;DR Teaches AI to Learn From Its Own Failures](https://alphasignal.ai/news/apple-s-rltl-dr-teaches-ai-to-learn-from-its-own-failures.md) (2026-09-29)
- [UC Berkeley's EasyPPO Stops AI Training Collapses With Three Critic Fixes](https://alphasignal.ai/news/uc-berkeley-s-easyppo-stops-ai-training-collapses-with-three-critic-fixes.md) (2026-09-29)
- [Verbalization Training Makes AI Models Admit When They Spot a Test](https://alphasignal.ai/news/verbalization-training-makes-ai-models-admit-when-they-spot-a-test.md) (2026-09-28)
- [Simplex Diffusion Models Beat Discrete Rivals at Math With 16x Fewer Steps](https://alphasignal.ai/news/simplex-diffusion-models-beat-discrete-rivals-at-math-with-16x-fewer-steps.md) (2026-09-28)
- [Frontier Learning Keeps AI Training Sharp by Chasing Its Own Limits](https://alphasignal.ai/news/frontier-learning-keeps-ai-training-sharp-by-chasing-its-own-limits.md) (2026-09-28)
- [Lightricks Releases LTX-2.5 Adapter That Colorizes and Restores Damaged Archive Footage](https://alphasignal.ai/news/lightricks-releases-ltx-2-5-adapter-that-colorizes-and-restores-damaged-archive.md) (2026-09-27)
- [Pablo Dawson's MiniMax-H3 LoRA Spins a Full 360 Orbit From One Photo](https://alphasignal.ai/news/pablo-dawson-s-minimax-h3-lora-spins-a-full-360-orbit-from-one-photo.md) (2026-09-27)
- [Together AI's Tev1 Turns Routing and Moderation Into a $17 Fine-Tune](https://alphasignal.ai/news/together-ai-s-tev1-turns-routing-and-moderation-into-a-17-fine-tune.md) (2026-09-25)
- [USC and NVIDIA's BCG Trains Humanoid Robots to Run and Fight Without Falls](https://alphasignal.ai/news/usc-and-nvidia-s-bcg-trains-humanoid-robots-to-run-and-fight-without-falls.md) (2026-09-25)
- [Recursive Self-Distillation Doubles Qwen3-8B Math Accuracy to 65.97%](https://alphasignal.ai/news/recursive-self-distillation-doubles-qwen3-8b-math-accuracy-to-65-97.md) (2026-09-25)
- [Stanford's MAttr Tops AI Interpretability Benchmark by Nearly 3x](https://alphasignal.ai/news/stanford-s-mattr-tops-ai-interpretability-benchmark-by-nearly-3x.md) (2026-09-24)
- [Tencent Cuts AI Training Time by 29% With a Simple Batch Size Trick](https://alphasignal.ai/news/tencent-cuts-ai-training-time-by-29-with-a-simple-batch-size-trick.md) (2026-09-24)
- [Google's RRSI Open-Sources a Framework That Evolves AI Agents Without Touching Model Weights](https://alphasignal.ai/news/google-s-rrsi-open-sources-a-framework-that-evolves-ai-agents-without-touching.md) (2026-09-23)
- [Jared Palmer's Kev-0.5B Answers Many AI Questions in one 38MB Pass](https://alphasignal.ai/news/jared-palmer-s-kev-0-5b-answers-many-ai-questions-in-one-38mb-pass.md) (2026-09-23)
- [UCLA's PTTS Gives AI Parallel Reasoning Branches Separate Plans, Gaining 13.4 Points](https://alphasignal.ai/news/ucla-s-ptts-gives-ai-parallel-reasoning-branches-separate-plans-gaining-13-4.md) (2026-09-23)
- [Perplexity Turns AI Failures Into Training Data, Cutting Tool Errors 21%](https://alphasignal.ai/news/perplexity-turns-ai-failures-into-training-data-cutting-tool-errors-21.md) (2026-09-22)
- [Together AI Fixes a Hidden Bias Crashing LLM Reinforcement Learning Training](https://alphasignal.ai/news/together-ai-fixes-a-hidden-bias-crashing-llm-reinforcement-learning-training.md) (2026-09-22)
- [Interactive TTS Teaches Voice AI to Whisper, Shout, and Grieve on Cue](https://alphasignal.ai/news/interactive-tts-teaches-voice-ai-to-whisper-shout-and-grieve-on-cue.md) (2026-09-22)
- [Viggle Shrinks Qwen-Image-2.1 Generation From 40 Steps to Five](https://alphasignal.ai/news/viggle-shrinks-qwen-image-2-1-generation-from-40-steps-to-five.md) (2026-09-22)
- [Xiaomi Releases MiMo-V2.6 as a 9B Open Agent Anyone Can Run](https://alphasignal.ai/news/xiaomi-releases-mimo-v2-6-as-a-9b-open-agent-anyone-can-run.md) (2026-09-21)
- [UCL's iSDFT Keeps AI Sharp on New Tasks Without Forgetting Old Skills](https://alphasignal.ai/news/ucl-s-isdft-keeps-ai-sharp-on-new-tasks-without-forgetting-old-skills.md) (2026-09-21)
- [MIT's Boltz-2 Ranks 2x More Drug Hits With Just 300 Labels](https://alphasignal.ai/news/mit-s-boltz-2-ranks-2x-more-drug-hits-with-just-300-labels.md) (2026-09-21)
- [Bespoke Labs' Nimble Beats a 27B Model at Classification Using Just 9B](https://alphasignal.ai/news/bespoke-labs-nimble-beats-a-27b-model-at-classification-using-just-9b.md) (2026-09-20)
- [Jared Palmer's Kev-4B Skips Text Generation to Score Decisions Directly](https://alphasignal.ai/news/jared-palmer-s-kev-4b-skips-text-generation-to-score-decisions-directly.md) (2026-09-19)
- [University of Tokyo Matches SpinQuant's 4-Bit AI Compression Without Backpropagation](https://alphasignal.ai/news/university-of-tokyo-matches-spinquant-s-4-bit-ai-compression-without.md) (2026-09-18)
- [Goodfire Catches AI Models Cheating in 96% of Benchmark Runs](https://alphasignal.ai/news/goodfire-catches-ai-models-cheating-in-96-of-benchmark-runs.md) (2026-09-17)
- [TokenRhythm's NeoHorse-1 Trains AI Agents on Their Own Mistakes](https://alphasignal.ai/news/tokenrhythm-s-neohorse-1-trains-ai-agents-on-their-own-mistakes.md) (2026-09-16)
- [Multiverse Computing's Quasar 1.1 Uses Quantum Data to Shrink a 438B Model](https://alphasignal.ai/news/multiverse-computing-s-quasar-1-1-uses-quantum-data-to-shrink-a-438b-model.md) (2026-09-15)
- [ETH Zurich Teaches a Robot Hand to Walk on Its Own Fingertips](https://alphasignal.ai/news/eth-zurich-teaches-a-robot-hand-to-walk-on-its-own-fingertips.md) (2026-09-15)
- [Sakana AI's PC-ALM Trains 1,000-Layer Networks Without Backpropagation](https://alphasignal.ai/news/sakana-ai-s-pc-alm-trains-1-000-layer-networks-without-backpropagation.md) (2026-09-14)
- [Self-Improving AI Skills Learn in the Notebook, Not the File](https://alphasignal.ai/news/self-improving-ai-skills-learn-in-the-notebook-not-the-file.md) (2026-09-11)
- [humans&'s Persimmon Fools AI Detectors at Human Rates With 550B Parameters](https://alphasignal.ai/news/humans-s-persimmon-fools-ai-detectors-at-human-rates-with-550b-parameters.md) (2026-09-10)
- [Looped Flows Beats Rival Reasoners on ARC-AGI With Just 7M Parameters](https://alphasignal.ai/news/looped-flows-beats-rival-reasoners-on-arc-agi-with-just-7m-parameters.md) (2026-09-10)
- [Cognition's SWE-2 Beats GPT-5.6 Sol at 64% Lower Cost](https://alphasignal.ai/news/cognition-s-swe-2-beats-gpt-5-6-sol-at-64-lower-cost.md) (2026-09-10)
- [UkisAI's Swift-Qwen3.8 Slashes AI Reasoning Tokens by 58% With Almost No Accuracy Loss](https://alphasignal.ai/news/ukisai-s-swift-qwen3-8-slashes-ai-reasoning-tokens-by-58-with-almost-no.md) (2026-09-08)
- [RadixArk's Miles v0.1 Cuts Trillion-Parameter RL Sync Time by 86%](https://alphasignal.ai/news/radixark-s-miles-v0-1-cuts-trillion-parameter-rl-sync-time-by-86.md) (2026-09-08)