Jensen Huang is out. What did China want from him?
Today, we’re talking about the Jensen Huang scandal at Nvidia, which is developing so quickly that we had to do an emergency update of this…
en
DeepSeek
138
15
HandsOnAid: Building and Evaluating an Offline First-Aid Assistant for Emergency Response Scenarios
Emergencies can occur where there is no phone signal and no ambulance nearby. A small language model on a phone can be helpful in such situations, but its advice needs to be safe and accurate. We built HandsOnAid, a 4-bit Gemma 2B model fine-tuned with QLoRA on FirstAidQA dataset, a corpus of 5,550 first-aid question-answer pairs. We compared seven configurations, including the base, 4-bit and 8-bit fine-tuned models, and versions with a length floor, safety gate, or retrieval. We used a 41-pair question-answer bank with medically reviewed references that followed ANZCOR guidelines, and a strict 0-5 score rubric, with twelve safety overrides. Four frontier LLM judges (GPT-5.6 Sol, Claude Opus 5, DeepSeek V4 Pro, GLM-5.3) blindly scored every answer, with 45 hidden controls used to check judge reliability. Fine-tuning improved the mean score by +1.08 over the base model to 2.34. The best configuration scored 2.51 out of 5 which was the fine-tuned 4-bit model with retrieval. A medical professional also conducted a blind evaluation of 40 answers across the base, the 4-bit fine-tuned, and the retrieval configurations, and the scores were similar to the LLM judges' average scores. A keyword audit found limited coverage of several details of ANZCOR guidelines in the training corpus, that are required by the rubric, which suggests that missing information in the training data could be a cause of the low scores. On a smartphone device (Pixel 7 Pro) with six CPU threads and no network connection, the 4-bit fine-tuned model produced its first displayable text in 6.86 seconds and a complete answer in 14.88 seconds on average over ten questions.
en
Can KV Cache Be Reused Across Model Updates? A DeepSeek-V4-Flash Case Study
Reusing key--value (KV) caches across language model updates could avoid repeated computation of shared context, but changes in model weights raise a fundamental question about what these caches preserve. We investigate direct cache reuse across four model updates, with a detailed study of DeepSeek-V4-Flash to Flash-0731. Our central finding is that caches can retain useful task information without faithfully reproducing the updated model's behavior. Retrieval accuracy varies substantially across updates, ranging from near-native performance to large losses. With source caches already available, reuse achieves median time-to-first-token speedups of 2.10--5.08on three longer-prefix QA benchmarks. To characterize the behavioral differences behind these outcomes, we introduce Functional Cache Divergence, which compares next-token distributions under reused and native caches. Direct KV cache reuse better preserves native model behavior in natural-text continuation, whereas compatibility degrades with increasing prefix length in single-token retrieval tasks. Answer-margin analyses help explain how correct retrieval decisions survive substantial internal state changes. Yet even agreement on the first generated token does not ensure agreement in answer content or stopping behavior. These findings establish task utility and fidelity to native inference as distinct criteria for evaluating cross-version KV cache reuse.
en
SLIC: Reinforcement Fine-Tuning Small LMs for Multi-Turn Analog Circuit Optimization
Analog integrated circuit (IC) design is a multi-turn, simulator-grounded engineering task that demands iterative reasoning over coupled performance tradeoffs, yet proprietary process design kits (PDKs) make hosted large language model workflows difficult to deploy in practical design settings. We present SLIC (Small LM Iterating on Circuits), a reinforcement fine-tuning framework that trains a 4B-scale, locally deployable language model for closed-loop analog circuit optimization. SLIC pairs structured circuit representations with discretized relative-update actions, domain- and success-gated multiobjective rewards, and Reward-Guided Rollout Steering (RGRS), which turns per-step simulator feedback into signals for both policy learning and online rollout construction in this sparse, highly constrained domain. We further release an open, goal-conditioned benchmark of 41 amplifier topologies with standardized protocols for in-distribution validation, held-out topology evaluation, and cross-technology-node transfer. Trained solely with open source 65 nm CMOS predictive technology models (PTM), SLIC-Random achieves 87.1% in-distribution four-specification Pass@1 within eight turns, surpassing the 763B DeepSeek-V4.1-Flash Think by 48.5 percentage points. Beyond the training distribution, SLIC-Softmax reaches 72.6% Pass@1 on held-out topologies, and the trained policies consistently transfer across five unseen technology nodes without additional fine-tuning. This transfer extends to a commercial TSMC 65 nm CMOS PDK, where SLIC-Random achieves 49.9% held-out-topology Pass@1, demonstrating practical deployment through local inference without transmitting proprietary data to external LLM services. Performance further scales with test-time simulator interactions, reaching up to 83.2% held-out-topology Pass@1 at a 32-turn budget.
en
Placing LLM Lookup Tables by Per-Layer Prefetch Deadlines
DeepSeek-V4.1's 203 GB of Engram n-gram tables sit in host DRAM today, taking 190 of the 223 GiB a 4-GPU serving process uses. Row indices depend only on input tokens, so reads can start with each step and overlap earlier layers. V4.1's two Engram layers have windows that differ by 30x (0.3 vs. 9.3 ms). We show that this deadline, together with session reuse and contention on a shared device, decides which tier each table can use and which layers need cache, and we give a model that predicts the resulting slowdown from measured read times. Agent and chat traffic reuse rows within sessions, and static hot sets transfer poorly. We build SlackTier, a table tier for SGLang that issues all lookups at step start. It reads the short-window layer first and splits a small GPU row cache across layers. The path runs inside the decode CUDA graph and is bit-exact with the original gather. Layer 14's table can go to NVMe uncached with no measurable decode cost. With both tables on four NVMe disks, mean time per output token on agent and chat replays rises at most 1.4% over unmodified SGLang, and host memory falls from 212 to 34 GiB. Both tables need no cache when read over RDMA (remote direct memory access) from another node's memory; one copy serves two instances. On a KV-constrained agent workload, using the freed memory as host KV cache at the same memory budget raises median throughput by 36% and cuts median p90 time to first token from 38 to 24 s.
en
Looping Deeper Without Growing the KV Cache: Recurrent Depth for Causal Encoder-Decoders
Looped language models deepen computation without adding parameters, yet caching full-context key-value (KV) states across layers and rounds severely bounds serving batch size. We design a looped causal encoder-decoder (CED) that achieves recurrent depth with zero round-wise KV growth, addressing where and how to loop an asymmetric stack in which only sparse encoder layers generate global KV. We find that recurrence must be placed in the encoder to iteratively refine global KV representations rather than re-reading static decoder projections, which both drives language-modeling gains and preserves long-context scaling. Furthermore, training the encoder loop under stochastic recurrence depth induces a fixed-point equilibrium where every round losslessly shares the last round's KV cache. Compounding cross-layer and cross-round KV sharing, the resulting Looped CED matches an unlooped CED's minimal memory footprint and recovers 86-93% of a looped Transformer's gain with 1/6 of its global KV. To test serving at production scale, we loop the 552B-parameter DeepSeek-V4.1-Flash for 8 rounds on one 4xGB200 node: with per-layer and per-round KV the loop cannot fit a single 128K-token request, whereas Looped CED holds 36.8x less KV per token, fits 38 such requests (as many as the unlooped model), and serves 3.6x higher throughput on 8K-token prompts.
en
Run DeepSeek R1 on Just 480GB of VRAM
DeepSeek R1 remains one of the top open-source models. This post shows how to run it efficiently on just 480GB of VRAM without sacrificing performance.
en
Agentic-OPD: Recovering Forgotten Agent Capabilities via On-Policy Distillation for Agentic Tasks
Reinforcement learning is now the mainstream paradigm for training LLMs on agentic tasks, yet standard practice keeps only the best-validation checkpoint and discards the rest. We show that this is lossy: in GRPO training of LLM agents, the best-validation checkpoint fails on problems that earlier checkpoints from the same run had already solved, and the union of problems solved across checkpoints far exceeds any single one. The discarded checkpoints thus form a teacher pool obtained at no additional training cost, which on-policy distillation (OPD) can mine to recover the lost coverage. Mining this pool raises two challenges: i) which checkpoint to distill from, as no checkpoint dominates the others and complementarity to the student is problem-specific; ii) where to trust it, as even a well-chosen teacher is only locally superior, so indiscriminate distillation overwrites behavior the student has already mastered. We propose Agentic-OPD, which treats the best-validation checkpoint as the student and all saved checkpoints as frozen teacher candidates, using the student's evolving capability gaps to guide both teacher selection and supervision filtering: the HCN policy (Highest Coverage Next) dynamically selects teachers by weighted coverage of the student's unsolved problems, while the LAD module (Locality-Aware Distillation) confines transfer to failed student rollouts on problems where the teacher performs better. Across five agentic task domains and 24 benchmarks and task subsets, Agentic-OPD improves over its initialization checkpoint by up to 7.4 points, with a 4B student surpassing the latest frontier models GPT-5.6 and DeepSeek-V4.1 on deep search. Agentic-OPD recasts a finished RL run as a reusable pool of policies and establishes post-RL capability recovery as a stage orthogonal to improvements in the RL algorithm itself.
en
BiVerify: Closing the Loop Between Backward and Forward Verification for Mathematical Self-Correction
Large language models (LLMs) have made substantial progress in mathematical reasoning but still produce incorrect solutions to complex problems, indicating the need for more reliable self-correction mechanisms. Mathematical self-correction typically identifies errors through verification, which can proceed backward or forward. The two directions provide complementary verification capabilities. Backward verification checks a candidate answer against the original constraints, bypassing complex intermediate reasoning, but cannot ensure correct reasoning or objective completeness. Forward verification follows the problem conditions and current solution, enabling it to locate reasoning errors and assess objective completeness, but may overlook key constraints in complex solution chains. We propose BiVerify, a dynamically scheduled framework for bidirectional mathematical verification. BiVerify begins with backward verification to screen the answer against the original constraints, followed by forward verification of the reasoning process; subsequent direction switching, state resetting, and re-verification are governed by verification outcomes. Upon failure, the system revises the candidate or verification flow until the same candidate passes both directions. We also introduce shared mathematical tools that separate logical reasoning from precise computation. We systematically evaluate BiVerify on five challenging mathematical reasoning benchmarks using multiple mainstream DeepSeek, GPT, and Gemini models. BiVerify exceeds the strongest representative public baseline evaluated in this work by 13.5 percentage points in average accuracy, with the margin reaching 31.2 percentage points on GPT. Further analysis shows that BiVerify's overall performance exceeds that of either verification direction alone.
en
AutoC2V: Synthesis-Guided LLM Training for QoR-Optimized C++-to-Verilog Generation
High-level synthesis (HLS) can translate C/C++ kernels into register-transfer level (RTL) designs, yet generating high-quality hardware still relies heavily on manual pragma tuning and repeated synthesis evaluation. We propose AutoC2V, a C++-to-Verilog generation framework driven by real synthesis feedback that enables large language models (LLMs) to directly generate RTL with improved quality of results (QoR) from synthesizable C++ kernels. We construct training data from real synthesis results, including synthesizable C++ source code, the corresponding Verilog RTL, QoR metrics, and optimization objectives. Using this dataset, we align the model toward synthesis-aware RTL generation through Supervised Fine-Tuning (SFT) and Group Relative Policy Optimization (GRPO). To reduce the high cost of repeatedly invoking real synthesis during GRPO, we design the QoR Prediction Reward Model (QoR RM), which predicts latency and resource metrics by encoding RTL into long-context semantic representations, thereby providing low-cost and continuous QoR reward signals for policy optimization. Experimental results demonstrate that AutoC2V not only substantially improves RTL functional pass rates, but also achieves superior latency and resource optimization compared to the unfine-tuned base model, representative HLS optimization approaches (HLSPilot, LLM-DSE, and AutoDSE), and advanced general-purpose commercial models (Claude Opus 4.8, GPT-5.5, Gemini 3.1 Pro, DeepSeek v4 Pro, and GLM-5.2). These results indicate that combining real synthesis data with efficient QoR reward modeling can further improve the ability of LLMs to generate high-quality RTL.
en
xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic
New text-generation model. Tags: qwen3_5, image-text-to-text, text-generation-inference, unsloth, reasoning, distillation, deepseek, sft
en♥ 117
RidgeFill: Training-Free Late-Layer Cache Fill for Faster Prefill
RidgeFill shortens prefill by not computing most of a model's upper layers for the prompt. The lower layers run exactly up to a split; from the split layer's state a per-layer map predicts every later layer's input, and each late layer writes its cache from that prediction. The design follows YOCO and DeepSeek-V4.1-Flash, where the projection is trained into the model. RidgeFill retrofits it onto models that were never trained for it: the map is a ridge regression fitted once in closed form from one pass over calibration text, the layer's own projections run unchanged, and a chosen set of rows, the last few thousand tokens, and every decoding step run the full model. Nothing is trained and no token is dropped. Most of the work was in measurement, finding a repeatable methodology. Loss averaged over all tokens hides RidgeFill's error. Needle retrieval in the filled region, loss on tokens that depend on distant context, and splice tests that swap filled keys or values for exact ones show which component carries it: the keys on Llama-3.1-8B, half of it in their scale, and the values and the recurrent state on Qwen3.5-9B. On Qwen3.8-27B in a llama.cpp fork on an A100, time to first token is 1.25 to 1.31x faster at 16K to 120K tokens with the maps in VRAM; RULER changes by -1.2 and -1.4 points and LongBench v2 by -1.4; the 95 % intervals include zero but do not rule out losses of a few points. On Qwen3.8-Flash-Next in vLLM it is 1.50 to 1.60x faster than exact at 9K to 32K tokens, with RULER's needle tasks at 99.8 against 100.0 and perplexity on novel tokens 2.65 % higher [-0.02 %, +4.93 %]. On three 8 to 9B reference models, at the same split and tail as the identity fill, the fitted map retrieves 49.1 and 49.6 more filled-region RULER points on the two dense models (at that split Llama's loss is worse), and its long-document loss is +/-0.03 nats of the released SwiftKV checkpoint's.
en
Unpacking the Compute Efficiency of Strong-to-Weak OPD against GRPO
On-policy distillation (OPD) trains a student on its own rollouts under a frozen teacher's per-token supervision. Does this dense signal reduce the training FLOPs needed to reach a target accuracy compared with reinforcement learning? We systematically compare vanilla reverse-KL OPD with Group Relative Policy Optimization (GRPO) across task difficulty, student and teacher size, domain, and optimization choices, measuring held-out accuracy against training FLOPs. With an external post-trained teacher, OPD generally loses to GRPO in accuracy per training FLOP at matched settings, even when the teacher is further trained with GRPO, and often suffers response-length inflation. Yet a different teacher changes the picture: when the teacher is the student's own GRPO checkpoint, OPD recovers that checkpoint's accuracy on math with 50fewer distillation FLOPs than GRPO needs to reach the same accuracy. Following the same teacher setting, we consistently observe this high efficiency with a larger student, on code, and in a second model family. On a single domain, however, the GRPO checkpoint is already the desired model, so distilling it adds to total compute. Across domains, separate GRPO checkpoints hold different capabilities, and no single checkpoint serves every domain. Multi-teacher OPD (DeepSeek-AI, 2026) consolidates per-domain GRPO experts into a single student that matches each expert's accuracy, with the distillation stage costing only 5.2% (on 2 domains) and 8.0% (on 4 domains) of the specialist training FLOPs. Counting specialist training and distillation together, this pipeline reaches joint GRPO's performance on the 2-domain mix with 42.0% fewer total FLOPs.
en
GC-MoE: Overlapping Expert-Parallel Token Dispatch with FP8/FP4 Grouped GEMM via CUDA Green Contexts
Under expert parallelism, mixture-of-experts (MoE) layers pay all-to-all token dispatch, grouped GEMMs, and a combine step in every layer. In decode each rank holds only 1-256 tokens: dispatch latency and weight traffic dominate while tensor-core utilization is inherently low. In fused operators such as DeepGEMM's Mega MoE, dispatch and computation overlap only by sharing every SM: dispatch warps compete with TMA/MMA and epilogue warps for registers, shared memory, and issue slots; every phase shares one CTA resource budget, synchronization spans the whole SM array, and every replay ends with a workspace-clean pass and a cross-rank barrier. We present GC-MoE, an open-source FP8/FP4 MoE inference operator for NVIDIA Hopper and Blackwell GPUs that uses CUDA Green Contexts to split the device into a small communication partition and a large compute partition. Each role gets its own kernel, resource budget, and synchronization domain, so dispatch and GEMM tune independently. A three-node CUDA Graph (four on Hopper with fused shared experts) launches the dispatch and grouped-GEMM kernels as concurrent roots; a token-ring handshake starts computation on the first filled ring blocks, and double-buffered receive counters remove the trailing cross-rank barrier between replays. The compute partition runs operand-swapped tcgen05 block-scaled MMA on Blackwell and register-A WGMMA with in-register FP4 decoding on Hopper, with fused SwiGLU, requantization, and optional shared experts. On 8 H20 GPUs GC-MoE accelerates DeepSeek-V4.1-Flash decode by 1.188x on average on both the routed and shared-expert suites (peaks 1.296x and 1.353x); on 8 B300 GPUs it accelerates GLM-5.2 and DeepSeek-V4.1-Flash by 1.106x (1.123x with fused shared experts, peak 1.209x), beating Mega MoE at all 54 measured points at matched accuracy. GC-MoE reserves only 20 of 148 SMs on B300 for dispatch.
en
Build efficient agents with knowledge distillation on Amazon SageMaker AI
Agentic assistants built on frontier models run a multi-step tool loop, and paying frontier prices on every step is hard to sustain at production scale. This post shows how to use sequence-level knowledge distillation (SeqKD) to transfer a frontier teacher's agentic behavior into a small, lower-cost student on Amazon SageMaker AI.
en
OBLITERATUS/DeepSeek-V4.1-Flash-1017-obliterated
New text-generation model. Tags: deepseek_v41_text, text-generation, obliteratus, abliterated, deepseek, moe, conversational, base_model:deepseek-ai/DeepSeek-V4.1-Flash-1017
nl♥ 112
nvidia/DeepSeek-V4.1-Flash-1017-NVFP4
New image-text-to-text model. Tags: Model Optimizer, deepseek_v41, nvidia, ModelOpt, DeepSeekV4, quantized, FP4, fp4
nl♥ 105