https://swastikroy.me 2026-06-20T22:14:52.616Z https://swastikroy.me/blog 2026-06-20T22:14:52.616Z https://swastikroy.me/about 2026-06-20T22:14:52.616Z https://swastikroy.me/publications 2026-06-20T22:14:52.616Z https://swastikroy.me/cv 2026-06-20T22:14:52.616Z https://swastikroy.me/blog/bradley-terry-elo 2026-06-20 https://swastikroy.me/blog/low-rank-matrix-approximation 2026-06-20 https://swastikroy.me/blog/svd-singular-value-decomposition 2026-06-20 https://swastikroy.me/blog/interp-attribution-methods 2025-06-20 https://swastikroy.me/blog/interp-circuits 2025-06-20 https://swastikroy.me/blog/interp-logit-lens 2025-06-20 https://swastikroy.me/blog/interp-open-problems 2025-06-20 https://swastikroy.me/blog/interp-representation-geometry 2025-06-20 https://swastikroy.me/blog/interp-sparse-autoencoders 2025-06-20 https://swastikroy.me/blog/interp-transformer-components 2025-06-20 https://swastikroy.me/blog/activation-steering 2025-02-02 https://swastikroy.me/blog/architecture-playground-transformer 2024-06-19 https://swastikroy.me/blog/diffusion-cfg 2024-06-19 https://swastikroy.me/blog/diffusion-ddim 2024-06-19 https://swastikroy.me/blog/diffusion-ddpm 2024-06-19 https://swastikroy.me/blog/diffusion-dit 2024-06-19 https://swastikroy.me/blog/diffusion-flow-matching 2024-06-19 https://swastikroy.me/blog/diffusion-latent 2024-06-19 https://swastikroy.me/blog/diffusion-score-matching 2024-06-19 https://swastikroy.me/blog/dropout-in-llms 2024-06-19 https://swastikroy.me/blog/grpo-vs-ppo 2024-06-19 https://swastikroy.me/blog/inference-continuous-batching 2024-06-19 https://swastikroy.me/blog/inference-memory-kv-quantization 2024-06-19 https://swastikroy.me/blog/inference-prefill-decode 2024-06-19 https://swastikroy.me/blog/inference-serving-at-scale 2024-06-19 https://swastikroy.me/blog/kv-cache-lifecycle 2024-06-19 https://swastikroy.me/blog/large-model-architecture-comparison 2024-06-19 https://swastikroy.me/blog/llm-alignment-pipeline 2024-06-19 https://swastikroy.me/blog/log-likelihood-to-perplexity 2024-06-19 https://swastikroy.me/blog/math-llm-activation-functions 2024-06-19 https://swastikroy.me/blog/math-llm-evaluation-metrics 2024-06-19 https://swastikroy.me/blog/math-llm-information-theory 2024-06-19 https://swastikroy.me/blog/math-llm-linear-algebra 2024-06-19 https://swastikroy.me/blog/math-llm-loss-functions 2024-06-19 https://swastikroy.me/blog/math-llm-optimization 2024-06-19 https://swastikroy.me/blog/math-llm-probability 2024-06-19 https://swastikroy.me/blog/math-llm-putting-it-together 2024-06-19 https://swastikroy.me/blog/math-llm-statistical-foundations 2024-06-19 https://swastikroy.me/blog/math-llm-svd-lora 2024-06-19 https://swastikroy.me/blog/padding-vs-packing 2024-06-19 https://swastikroy.me/blog/ppo-full-training-step 2024-06-19 https://swastikroy.me/blog/ppo-loss-per-token 2024-06-19 https://swastikroy.me/blog/rai-privacy-leakage 2024-06-19 https://swastikroy.me/blog/rai-red-teaming-vs-automated-eval 2024-06-19 https://swastikroy.me/blog/rai-runtime-guardrails 2024-06-19 https://swastikroy.me/blog/rai-safety-benchmark-design 2024-06-19 https://swastikroy.me/blog/rl-for-llms-agentic-systems 2024-06-19 https://swastikroy.me/blog/rl-for-llms-learning-from-feedback 2024-06-19 https://swastikroy.me/blog/rl-for-llms-policy-gradients 2024-06-19 https://swastikroy.me/blog/rl-for-llms-skill-acquisition 2024-06-19 https://swastikroy.me/blog/rl-for-llms-why-rl 2024-06-19 https://swastikroy.me/blog/ssl-contrastive-learning 2024-06-19 https://swastikroy.me/blog/ssl-ijepa 2024-06-19 https://swastikroy.me/blog/ssl-jepa-idea 2024-06-19 https://swastikroy.me/blog/ssl-masked-autoencoders 2024-06-19 https://swastikroy.me/blog/ssl-vjepa 2024-06-19 https://swastikroy.me/blog/ssl-why-self-supervised 2024-06-19 https://swastikroy.me/blog/ssl-world-models-vision 2024-06-19 https://swastikroy.me/blog/synthetic-data-for-alignment 2024-06-19 https://swastikroy.me/blog/transformer-evolution-gpt2 2024-06-19 https://swastikroy.me/blog/transformer-training-architecture-internals 2024-06-19 https://swastikroy.me/blog/transformer-training-attention-variants 2024-06-19 https://swastikroy.me/blog/transformer-training-data-efficiency 2024-06-19 https://swastikroy.me/blog/transformer-training-debugging 2024-06-19 https://swastikroy.me/blog/transformer-training-normalization 2024-06-19 https://swastikroy.me/blog/transformer-training-optimizer 2024-06-19 https://swastikroy.me/blog/transformer-training-positional-encoding 2024-06-19 https://swastikroy.me/tags/activation-functions 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/activation-patching 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/activation-steering 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/adam 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/agents 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/ai-safety 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/alignment 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/architecture 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/attention 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/attention-heads 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/attribution 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/batching 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/bayesian 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/benchmarking 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/circuits 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/classifier-free-guidance 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/code 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/code-generation 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/comparison 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/conditioning 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/contrastive-learning 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/data 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/data-curation 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/ddim 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/ddpm 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/debugging 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/dictionary-learning 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/diffusion 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/dimensionality-reduction 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/dino 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/direct-logit-attribution 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/dit 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/elo 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/embodied-ai 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/entropy 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/evaluation 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/features 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/fine-tuning 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/flow-matching 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/generative-models 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/gpt 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/gqa 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/gradient-descent 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/grpo 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/guardrails 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/history 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/i-jepa 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/image-generation 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/indirect-object-identification 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/inference 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/information-theory 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/integrated-gradients 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/interactive 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/interpretability 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/jepa 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/kv-cache 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/language-models 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/latency 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/latent-diffusion 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/lime 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/linear-algebra 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/linear-representation-hypothesis 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/llm 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/llm-training 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/logit-lens 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/long-context 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/lora 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/loss-functions 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/low-rank 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/mae 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/masked-autoencoders 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/math 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/mechanistic-interpretability 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/memorization 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/memory 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/metrics 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/mlp-layers 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/model-compression 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/moe 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/neural-networks 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/normalization 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/open-problems 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/optimization 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/optimizer 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/pca 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/perplexity 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/planning 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/policy-gradient 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/positional-encoding 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/ppo 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/preference-learning 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/pretraining 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/privacy 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/probability 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/production 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/qlora 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/quantization 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/ranking 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/reasoning 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/rectified-flow 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/red-teaming 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/regularization 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/representation-learning 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/residual-stream 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/responsible-ai 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/reward-model 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/reward-models 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/rl 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/rlhf 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/rope 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/safety 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/saliency 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/sampling 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/score-matching 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/self-supervised 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/sequence-packing 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/serving 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/sft 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/shap 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/sparse-autoencoders 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/ssl 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/stable-diffusion 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/statistics 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/superposition 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/svd 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/swiglu 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/synthesis 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/synthetic-data 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/systems 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/throughput 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/training 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/training-dynamics 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/transformers 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/v-jepa 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/vae 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/video 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/vision-transformers 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/vllm 2026-06-20T22:14:52.917Z https://swastikroy.me/tags/world-models 2026-06-20T22:14:52.917Z