feat(wip): forgetting lit explorer module + knowledge base

fix: remove d_horizon tautology + audit-driven correctness/honesty fixes + test net

docs: refresh hf-space-readme (25 modes) + add v0.9 announcement draft

docs: update README/Space card for v0.8–v0.9 (25 modes, 3 new tools)

fix: footer "Paper repo" pointed to a 404 (github.com/karlesmarin/NeurIPS)

v0.9.4: Launch-Flag Generator mode + Zenodo record update

v0.9.3: GGUF repo autocomplete auto-lists quant files on select

v0.9.2: HF autocomplete on GGUF repo input (gguf-tag filtered)

v0.9.1: GGUF Validity Bridge mode + binary header parser

v0.9.0: YaRN / RoPE context-extension planner mode

v0.8.10: fix community feed + falsification gated behind Pyodide; point LIMITATIONS link to GitHub-rendered

v0.8.9: make Validity Gate pill + Diagnostics auto-open reactive to γ_obs changes

v0.8.9: auto-open Diagnostics + summary pill when γ Validity Gate fires

v0.8.9: γ Validity Gate + LIMITATIONS.md regime-of-validity addendum

v0.8.8: wire HF autocomplete to longscore-input

v0.8.8: add LongScore to "What do you want to do?" entry tile

fix(v0.8.8): rename fmtPct → lsFmtPct in LongScore (collided with existing top-level fmtPct)

v0.8.8 LongScore mode — anti-bullshit pack #14 + Hub badge readability fix

hub: add multilingual_token_tax entry → 🌍 Multilingual Tokenizer Tax

v0.8.7 Multilingual Tokenizer Tax Calculator — anti-bullshit pack #13

v0.8.6+ niah: config.json fetch gets mirror fallback (gated → unsloth)

v0.8.6 NIAH RULER calibration — anti-bullshit pack #12

v0.8.5+ spec-decode: fetch via /resolve/main/ (LFS pointer fix)

v0.8.5+ spec-decode: open-mirror fallback for gated models (no token needed)

v0.8.5-fix: spec-decode examples → open-weight pairs (HF auth not required)

v0.8.5 Speculative-Decode Compatibility Checker — anti-bullshit pack #11

v0.8.4 Prompt-Cache Diff Predictor — anti-bullshit pack #10

v0.8.3 PEFT Anti-Pattern Checker — anti-bullshit pack #9

v0.8.2 JSON CoT-aware Linter — anti-bullshit pack #8

v0.8.1 Solutions Hub — integrator portal (30 pains × 65 external tools)

v0.8.0 Benchmark Saturation Detector — anti-bullshit pack #6

v0.7.7-fix: ⓘ tooltips on each task tile, 4 langs (EN/ES/FR/ZH)

fix(ux): collapse mode-section by default — task-tiles is now primary entry

v0.7.7: UX restructure — task tiles + 'Why static?' README FAQ

chore: gitignore draft post files (HF_FORUM_POST / HN_POST / EMAIL_*)

fix(readme): trim short_description to ≤60 chars (HF Space yaml limit)

docs: README v0.7 — anti-bullshit pack section + tags + punchier short_description

fix(niah): apply YaRN/rope_scaling factor to theta so extended-context models aren't false-broken

fix: attach HF autocomplete to niah-id input (was missed in v0.7.6)

v0.7.6: NIAH→reasoning predictor (anti-bullshit #7) + Drift CSS fix

v0.7.5: Cross-framework drift bound (anti-bullshit #6) + autocomplete UX

v0.7.4: HF Hub search-as-you-type autocomplete on all model id inputs

v0.7.3: Quant-regime classifier (anti-bullshit pack #5)

chore: triage untracked — gitignore scratch outputs, move helper script

v0.7.2: Arena CI + Contamination Prior + v0.7 Help/Inventory documentation

v0.7.1: Chat-template Sniffer (anti-bullshit #2) + 9th mode

v0.7.0: SWA Unmasker (anti-bullshit #1) + foldable main panels + preset auto-fill

v0.6.2: landing rework — minimalist hero + 3 modals (manual / quick start / inventory)

v0.6.1: 5 audit-driven gap fixes (privacy, exports, a11y, tooling)

v0.6: TAF Card hero+accordion UI + γ_check + Lean+Mathlib badges