WebBrain Compass VLD v6 (2B)
WebBrain Compass VLD v6 2B is WebBrain's default multimodal WebGPU in-browser browser agent model, succeeding Compass VLD Tiny v5.
It fine-tunes webbrain-one/lfm3-vld-2b-multipass (derived from LiquidAI/LFM3-VL-2B) on a license-clean 70k multimodal browser agent training mix, operating entirely without tool calls in the browser across three core roles:
- Tiny Ask & Standalone Chat: Single-turn and multi-turn Q&A from page text/screenshot and offline Wikipedia / Emergency Box passages with byte-exact citation tokens (
[WB-W-XXXXXX]). Politely refuses Act/Dev requests in user's language. - Vision Description: WebBrain standard six-section structured descriptions of browser viewports (Page purpose, Visible text, Inputs, State signals, Blockers, Unknowns).
- Decisions via Verify Protocol: Fast verify scoring ('Proposed answer: ... Is the proposed answer correct? Answer Yes or No.') evaluated via Yes-minus-No logit margins and linear probe on final hidden state (
decision_hidden).
Companion Repositories
webbrain-one/webbrain-compass-vld-v6-2b-onnx: WebGPU ONNX bundle (decoder_q8f16.onnx,decoder_q4f16.onnxwithdecision_hiddenexposed,vision_encoder_fp16.onnx, and trained linear decision probes).webbrain-one/webbrain-compass-vld-v6-2b-gguf: GGUF exports (BF16,Q8_0,Q4_K_M, andmmproj-F16) forllama.cpp(/v1/chat/completions+/v1/systemone).webbrain-one/webbrain-compass-vld-v6-2b-lora: LoRA adapter (r=32, alpha=64) and projector weights.
Evaluation Results & Gates
| Evaluation Gate | Target | Achieved Result | Status |
|---|---|---|---|
| Vision Suite (Q4_K_M) (E1) | >= 44/100 strict & > 58 (v5 baseline) | 62/100 strict (mean rubric: 0.8930, 0 errors, 1.13s latency) | PASS |
| Vision Suite (Q8_0) (E1) | Reference verification | 66/100 strict (mean rubric: 0.8994, 0 errors, 1.50s latency) | PASS |
| Six-Section Validity (E5) | >= 99.0% | 99.00% (495/500 held-out VL-2 val cases valid) | PASS |
| d1 final_test Raw Verify (E3) | >= 120/126 verify, 24/24 comp, 0/48 false comp, >= 23/24 unknown | 122/126 verify, 24/24 comp, 0/48 false comp, 24/24 unknown | PASS |
| In-Browser Chat Held-Out (E4) | Lang >= 95%, Act >= 95%, Citations 100%, 0 leaks | Lang: 95.95%, Act: 100.0%, Citations: 100.0%, Leaks: 0 | PASS |
| Routing Replay (E2) | Structured >= 95/100, loose > 32/89, 0 length hits | Structured: 96/100, Loose: 33/89, Length hits: 0 | PASS |
License & Provenance Audit
- Base Model:
LiquidAI/LFM3-VL-2B(LFM Open License v1.0). - Training Mix: 70,717 train rows, 1,631 val rows across 5 license-audited components:
vision_vl2(46,879): CC-BY-4.0 (WebSight, ScreenParse, MiniWoB++, synthetic web renders).vision_curated(5,880): Permissive (Orchard MIT, markov-ai Apache-2.0, look-before-you-click MIT).decision_d1(7,236): Synthetic / Internal d1 balanced-V3 browser decision tasks.routing_compass(6,950): Commercial manifest (WebBrain traces, Orchard MIT, Reduced-A11y CC BY 4.0, markov-ai Apache-2.0, Mind2Web CC BY 4.0).chat_inbrowser(3,822): CC-BY-4.0 / CC-BY-SA (Multilingual page QA, Turkish >= 20%, Act refusals, offline RAG, query translations).
- Strict Exclusions: WebChain (strictly excluded), WebLINX (strictly excluded), Compass v3 (strictly excluded).