Skip to content
tezvyn:

All bites

The whole library, newest first. Filter by what you are here for, or pick a topic if you already know.

4330 bites

Page 45

LLMs & Generative AI1 min read

How to evaluate a RAG system end to end

Measure retrieval (recall, precision, MRR, NDCG) and generation (faithfulness, answer relevance) separately, plus end-to-end correctness.

LLMs & Generative AI1 min read

How MMLU works and the contamination problem

MMLU is multiple-choice across 57 subjects scored by accuracy; contamination means test items leaked into pretraining, inflating scores.

LLMs & Generative AI1 min read

Reference-free evaluation for open-ended dialogue

ROUGE punishes valid paraphrases; use reference-free LLM-as-judge or learned scorers rating coherence, relevance, and groundedness.

LLMs & Generative AI1 min read

Setting up an LLM-as-a-judge evaluation

Define rubric, prefer pairwise comparison, randomize order, calibrate against humans; control position, verbosity, and self-preference bias.

LLMs & Generative AI1 min read

What RLHF is and the safety problem it solves

Collect human preference rankings, train a reward model, fine-tune the policy with PPO; it aligns outputs with human intent the loss function cannot specify.

LLMs & Generative AI1 min read

Red teaming LLMs and harmful output categories

Red teaming deliberately probes a model to elicit harmful behavior; categories include toxic or hateful content and dangerous instructions like weapons or malware.

LLMs & Generative AI1 min read

Designing input and output guardrails for a chatbot

Input guardrails filter or classify user prompts (injection, off-topic, PII) before the model; output guardrails validate responses for toxicity, leakage, and policy before sending.

LLMs & Generative AI1 min read

Prompt injection versus jailbreak, and defenses

Injection hijacks the model via untrusted data overriding developer instructions; jailbreak coaxes a model past its safety policy. Defense: separate trusted instructions from untrusted data and filter.

LLMs & Generative AI1 min read

Model quantization benefits and risks

Quantization stores weights and activations in lower precision like INT8 or INT4; benefits are smaller memory and faster, cheaper inference; risk is accuracy loss.

LLMs & Generative AI2 min read

How the KV cache speeds transformer generation

Cache stores past keys and values so each new token only computes its own K, Q, V instead of recomputing all prior tokens, cutting per-step cost from quadratic to linear.

LLMs & Generative AI1 min read

Teacher-student knowledge distillation

A small student learns to mimic a large teacher's soft probability outputs, not just hard labels; goal is a compact model retaining most capability.

LLMs & Generative AI1 min read

PTQ versus QAT for INT8 quantization

PTQ quantizes a trained model with light calibration, fast and cheap but more accuracy loss; QAT simulates quantization during training, higher accuracy but costly.

LLMs & Generative AI2 min read

What memory problem PagedAttention solves

Pre-allocating contiguous max-length cache per sequence wastes memory through internal and external fragmentation; PagedAttention stores KV in fixed non-contiguous blocks like OS paging.

LLMs & Generative AI2 min read

Tensor versus pipeline parallelism for large models

Tensor parallelism splits individual layers across GPUs needing fast interconnect; pipeline parallelism splits layers into stages across GPUs.

LLMs & Generative AI1 min read

Core insight behind GPTQ and AWQ

Not all weights matter equally; GPTQ minimizes layer output error using second-order info, AWQ protects salient weight channels tied to large activations.

LLMs & Generative AI1 min read

Rule-based versus model-based LLM guardrails

A guardrail is a programmatic check constraining LLM I/O; rule-based uses regex or blocklists, model-based uses a classifier like a moderation model to detect harmful content.

LLMs & Generative AI2 min read

Detecting RAG hallucinations with a confidence score

Decompose the answer into claims, verify each against retrieved context with NLI or an LLM judge, aggregate into a faithfulness confidence score, and flag unsupported claims.

LLMs & Generative AI2 min read

Direct versus indirect injection and agent defenses

Direct injection comes from the user prompt; indirect hides in third-party data the agent ingests like web pages.

LLMs & Generative AI1 min read

Differential privacy vs utility in LLM fine-tuning

Clipping plus calibrated noise per step, smaller epsilon means stronger privacy but degraded accuracy, tracking the privacy budget across epochs.

LLMs & Generative AI1 min read

Fine-tuning vs RAG for daily-updated docs

Choose RAG because docs change daily, embed and index chunks in a vector store, retrieve top matches and inject into the prompt.