designingintelligencedesigning intelligence
Models today are largely black boxes. We can prompt them, fine-tune them, and add guardrails, but our understanding of what they actually learned and why they behave the way they do remains limited.
attribution
Trace every response word back to the prompt tokens that caused it, and see how signal flows through layers to get there.
Causal graph of how signal flows through layers. Thicker edges carry more weight. See which layers matter for any output.
edge weight = causal signallogit lens
See what the model thinks at every layer as it builds toward a final answer. Watch a vague token sharpen into a confident prediction.
weight editing
Edit a weight directly. See the delta and the response change side by side. No retraining, no guessing — the diff is immediate and traceable.
diff
Connect any two checkpoints (base vs fine-tuned, fp16 vs int4) and see exactly which weights shifted, which features got compressed, and what behaviour changed.
benchmarks
Three suites built into Aquin. Run them on any checkpoint, edit, or quantization pass. Know immediately whether a change made the model better or worse.
EditBench
edit fidelitySurgical precision. Does the edit change only what you intended?
FineTuneDiff
checkpoint diffWhat actually changed between base and fine-tuned at the weight level.
InterpScore
interpretabilityHow cleanly do features map to human-readable concepts?
| run | EditBench | FineTuneDiff | InterpScore | delta |
|---|---|---|---|---|
| llama-3.2-1b · base | 71 | 64 | 59 | baseline |
| llama-3.2-1b · sft-v1 | 78 | 79 | 63 | +9 avg |
| llama-3.2-1b · sft-v2 | 82 | 83 | 70 | +5 avg |
| llama-3.2-1b · int4-quant | 74 | 71 | 61 | -9 avg |
| llama-3.2-1b · rome-edit-1 | 94 | 88 | 73 | +14 avg |
model support
Aquin works across the open-source ecosystem. Each model is validated against a case study before it ships.
agentic systems
Aquin closes the loop. Two autonomous agents: one that finds the weight, one that edits it. Both run without you touching a hyperparameter.
Observe a behaviour, trace it, plan a minimal edit, apply it, verify. Fully autonomous.
Describe a behaviour. The agent traces it to the exact circuit responsible and tells you whether it's a weight-level decision or a surface patch.
human readability
Model internals are not inherently unreadable. Aquin translates activations, weights, and layer states into language an engineer can reason about.
| weight | raw | label |
|---|---|---|
| L14 · MLP W_out [2048,11] | 0.847 | capital city associations |
| L8 · attn head 3 · V | -0.312 | geographic suppression |
| L12 · MLP W_in [512,2048] | 0.601 | factual recall trigger |
| L6 · attn head 7 · Q | 0.229 | question parsing |
factual checks
Most models ship as black boxes. You have no way to know what they learned to suppress, amplify, or distort. Aquin surfaces it.
Trace which features consistently skew outputs along political, demographic, or cultural lines. See the weight, not just the symptom.
Find what the model refuses to say and why. Identify suppression circuits. See whether refusals are weight-level decisions or surface-level RLHF patches.
isolated models
Every user gets their own model instance per tab. State never bleeds between sessions. Work on Llama in one tab, Mistral in another, no interference.
bulk editing
Apply a set of edits across multiple layers in a single operation. Queue them, run them, verify the aggregate delta.
aipedia
A living, community-indexed knowledge base of model features. Every behaviour, every circuit, every weight pattern. Searchable. Citable. Growing.
| feature | model | layer | circuit | confidence |
|---|---|---|---|---|
| capital city recall | Llama 3.2 1B | L14 | MLP W_out [2048,11] | 94% |
| hedging language | Llama 3.2 1B | L8 | attn head 3 · V | 87% |
| geographic association | Mistral 7B | L11 | MLP W_in [512,2048] | 81% |
| refusal circuit | Gemma 2B | L9 | attn head 7 · Q | 76% |
Not sure if Aquin is right for you?
Join ServerAquin