Encoder vs Decoder vs Encoder-Decoder Transformers

The Three Families

The original 2017 Transformer (“Attention Is All You Need”) was an encoder-decoder. The field then diverged into three distinct families, each optimised for different tasks.

1. Encoder-Only: BERT-style

Each encoder block contains:

• Bidirectional self-attention (every token sees every other token)
• Feed-forward network
• Layer norm + residual connections

Training objective: Masked Language Modelling (MLM). Random tokens in the input are replaced with [MASK], and the model predicts them. Because the answer is already in the sequence (just hidden), the model can attend bidirectionally.

What this is good at:

• Sentence classification (spam detection, sentiment)
• Token classification (NER, POS tagging)
• Question answering (span extraction)
• Sentence embeddings (semantic search)

What this cannot do: autoregressive generation. Generating token N+1 requires seeing token N+1 (bidirectional), which is circular during inference.

Examples: BERT, RoBERTa, DeBERTa, ALBERT, ModernBERT.

2. Decoder-Only: GPT-style

Each decoder block contains:

• Causal (masked) self-attention (each token sees only past tokens)
• Feed-forward network
• Layer norm + residual connections

Note: there is no cross-attention in decoder-only models. Each decoder block has only two sub-layers (not three), because there is no encoder output to attend to.

Training objective: Next-token prediction. Given tokens 1…N, predict token N+1. The causal mask ensures no peeking.

What this is good at:

• Text generation (stories, code, completions)
• In-context learning (few-shot prompting)
• Instruction following (with RLHF/fine-tuning)
• Anything you can frame as completion

What this is less natural for: tasks that require reading the full input before producing an output (e.g., translation, summarisation), though modern large decoder-only models handle these with prompting anyway.

Examples: GPT-2, GPT-3, GPT-4, LLaMA, Mistral, Gemma, Claude.

3. Encoder-Decoder: T5-style

The encoder processes the full input bidirectionally. The decoder generates the output token by token, using:

• Causal self-attention (on its own generated tokens so far)
• Cross-attention (queries the encoder’s output at each step)
• Feed-forward network

Training objective: Span corruption (T5) or similar sequence-to-sequence objectives.

What this is good at:

• Machine translation (full input available, output generated)
• Summarisation (read document, write summary)
• Question answering with generation (read context, write answer)
• Any task naturally framed as input → output transformation

Examples: T5, BART, mT5, Flan-T5, NLLB (translation).

Side-by-Side Comparison

The Attention Mask Differences

Encoder self-attention (bidirectional):
✓ ✓ ✓ ✓
✓ ✓ ✓ ✓
✓ ✓ ✓ ✓
✓ ✓ ✓ ✓

Decoder self-attention (causal):
✓ ✗ ✗ ✗
✓ ✓ ✗ ✗
✓ ✓ ✓ ✗
✓ ✓ ✓ ✓

Decoder cross-attention (full encoder access):
✓ ✓ ✓ ✓   ← decoder pos 1 attends to all encoder positions
✓ ✓ ✓ ✓   ← decoder pos 2 attends to all encoder positions

The mask tells the whole story. Encoder: open. Decoder: lower triangular. Cross-attention: open to the encoder.

Concrete Worked Example: Translating “The cat sat”

To make the three architectures concrete, consider translating “The cat sat” into French (“Le chat s’est assis”).

Encoder (reads input): Tokens [The, cat, sat] enter simultaneously. At layer 1, cat attends to both The and sat, bidirectional context tells it this is a subject noun, not a verb. All positions are processed in parallel.

Decoder step 1 (generates “Le”): Input to decoder: [<BOS>]. Causal self-attention: only position 0 is visible. Cross-attention: <BOS> queries the encoder’s full representation of [The, cat, sat] and retrieves a weighted mixture centred on “The”. Output distribution peaks at “Le”.

Decoder step 2 (generates “chat”): Input: [<BOS>, Le]. Causal self-attention: position 1 can see position 0 (“Le”) but not future tokens. Cross-attention: “Le” queries the encoder and attends heavily to “cat”. Output: “chat”.

Decoder step 3 (generates “s’est”): Input: [<BOS>, Le, chat]. Cross-attention now attends to the verb “sat”. Autoregressive chain continues until <EOS>.

This step-by-step shows why encoder-decoder wins for translation: the encoder reads all context before any generation begins, and cross-attention lets each decoder step query that full context freely.

Summary

The three architectures are not better or worse in absolute terms, they are optimised for different settings:

• Understanding a fixed input? → Encoder-only
• Generating open-ended text? → Decoder-only
• Transforming one sequence into another? → Encoder-decoder

Modern LLMs (GPT-4, Claude, LLaMA) are decoder-only, using scale and prompting to cover all three use cases. But for specialised tasks with a clear input-output structure and limited compute, encoder-decoder models remain competitive.

References

• Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention Is All You Need. NeurIPS 2017 (original encoder-decoder Transformer for machine translation).
• Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. NAACL 2019 (BERT: the canonical encoder-only Transformer for classification and understanding tasks).
• Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2020). Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer. JMLR 2020 (T5: unifies NLP tasks under a single encoder-decoder text-to-text format).
• Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., et al. (2020). Language Models are Few-Shot Learners. NeurIPS 2020 (GPT-3: decoder-only architecture at 175B parameters demonstrating few-shot learning).