{ "trained_at": "2026-04-27T06:37:23.331946+00:00", "data_run_date": "2026-04-25", "llm_run_date": "2026-04-26", "version": "v0.5-llm-features", "description": "v0.4 + LLM-derived features (name_kind, cultural_origin, fame_score, crypto_relevance, brand_collision_risk, description_emb_pca32)", "parent_version": "v0.4", "embedders": [ "mpnet-finetuned (from v0.3)", "BAAI/bge-base-en-v1.5" ], "splits": { "train": { "rows": 265240, "start": "2022-01-28", "end": "2023-09-30" }, "val": { "rows": 3545, "start": "2023-10-01", "end": "2023-12-31" }, "test": { "rows": 2744, "start": "2024-01-01", "end": "2024-05-04" } }, "feature_count": 212, "n_llm_features": 56, "feature_cols": [ "pca_000", "pca_001", "pca_002", "pca_003", "pca_004", "pca_005", "pca_006", "pca_007", "pca_008", "pca_009", "pca_010", "pca_011", "pca_012", "pca_013", "pca_014", "pca_015", "pca_016", "pca_017", "pca_018", "pca_019", "pca_020", "pca_021", "pca_022", "pca_023", "pca_024", "pca_025", "pca_026", "pca_027", "pca_028", "pca_029", "pca_030", "pca_031", "pca_032", "pca_033", "pca_034", "pca_035", "pca_036", "pca_037", "pca_038", "pca_039", "pca_040", "pca_041", "pca_042", "pca_043", "pca_044", "pca_045", "pca_046", "pca_047", "pca_048", "pca_049", "pca_050", "pca_051", "pca_052", "pca_053", "pca_054", "pca_055", "pca_056", "pca_057", "pca_058", "pca_059", "pca_060", "pca_061", "pca_062", "pca_063", "len", "n_digits", "n_letters", "n_special", "n_lower", "n_upper", "is_palindrome", "is_all_digits", "is_all_letters", "is_ascii", "has_unicode", "starts_digit", "ends_digit", "max_char_run", "n_unique_chars", "in_wikipedia", "in_geonames", "in_us_firstname", "in_iso3166", "in_ticker", "in_sec_edgar", "in_wiktionary_en", "wordlist_hits", "club__prepunk_full_rankings", "club__common_english", "club__common_animals", "club__gamertags", "club__firstnames_usa", "club__top_nouns", "club__us_government", "club__pokemon_gen1", "club__catholicism", "club__gamertags_double", "club__wikidata_top_fantasy_char", "club__social_handles", "club__pokemon_gen2", "club__top500_cities_global", "club__top500_cities_usa", "club__mythical_creatures", "club__conspiracy_theories", "club__crypto_terms", "club__natural_wonders", "club__pokemon_gen3", "club__familynames_usa", "club__country_codes", "club__sports", "club__top_crypto_names", "club__logistics", "club__home", "club__historic_figures", "club__holidays", "club__top_crypto_tickers", "club__paranormal", "club__performing_arts", "club__fine_art", "club__currency_symbols", "club__luxury", "club__personas", "club__finance_terms", "club__currency_names", "club__pokemon_gen4", "club__gen_alpha", "club__crayola_classic", "club__memes", "club__us_states", "n_clubs", "trademark_conflict", "name_age_days", "prior_transfer_count", "fg_value", "eth_tvl_usd", "eth_stable_mcap", "eth_dex_volume", "nft_total_fee_usd", "fame_score", "crypto_relevance_ord", "brand_collision_risk_ord", "kind__concept", "kind__random", "kind__brand", "kind__surname", "kind__first_name", "kind__abbreviation", "kind__place", "kind__other", "kind__unknown", "origin__english", "origin__none", "origin__mixed", "origin__spanish", "origin__german", "origin__french", "origin__japanese", "origin__chinese", "origin__italian", "origin__slavic", "origin__korean", "origin__arabic", "origin__other", "origin__unknown", "desc_pca_000", "desc_pca_001", "desc_pca_002", "desc_pca_003", "desc_pca_004", "desc_pca_005", "desc_pca_006", "desc_pca_007", "desc_pca_008", "desc_pca_009", "desc_pca_010", "desc_pca_011", "desc_pca_012", "desc_pca_013", "desc_pca_014", "desc_pca_015", "desc_pca_016", "desc_pca_017", "desc_pca_018", "desc_pca_019", "desc_pca_020", "desc_pca_021", "desc_pca_022", "desc_pca_023", "desc_pca_024", "desc_pca_025", "desc_pca_026", "desc_pca_027", "desc_pca_028", "desc_pca_029", "desc_pca_030", "desc_pca_031", "knnmp_count", "knnmp_mean_log", "knnmp_median_log", "knnmp_p90_log", "knnmp_max_sim", "knnmp_min_sim", "knnmp_log_max", "knnmp_log_min", "knnbg_count", "knnbg_mean_log", "knnbg_median_log", "knnbg_p90_log", "knnbg_max_sim", "knnbg_min_sim", "knnbg_log_max", "knnbg_log_min" ], "pca_dim_concat": 64, "pca_dim_description": 32, "name_kind_values": [ "concept", "random", "brand", "surname", "first_name", "abbreviation", "place", "other", "unknown" ], "cultural_origin_values": [ "english", "none", "mixed", "spanish", "german", "french", "japanese", "chinese", "italian", "slavic", "korean", "arabic", "other", "unknown" ], "best_iteration": 712, "xgb_params": { "objective": "reg:squarederror", "eval_metric": "rmse", "tree_method": "hist", "device": "cuda", "max_depth": 7, "learning_rate": 0.04, "subsample": 0.85, "colsample_bytree": 0.65, "min_child_weight": 8, "reg_alpha": 0.5, "reg_lambda": 2.0, "seed": 42 }, "metrics": { "train": { "r2_log": 0.8462194204330444, "rmse_log": 0.6332300305366516, "mae_log": 0.4247952699661255, "median_ape": 0.26942259073257446, "bias_log": -0.0002484263095539063 }, "val": { "r2_log": 0.6880427002906799, "rmse_log": 1.0231839418411255, "mae_log": 0.7111709713935852, "median_ape": 0.4590761661529541, "bias_log": 0.11442188918590546 }, "test": { "r2_log": 0.41624486446380615, "rmse_log": 1.420867681503296, "mae_log": 1.1492146253585815, "median_ape": 1.0128521919250488, "bias_log": 0.6260586380958557 } }, "top_features": [ { "name": "knnmp_mean_log", "gain": 2001.844970703125 }, { "name": "knnmp_median_log", "gain": 1155.0853271484375 }, { "name": "knnmp_p90_log", "gain": 832.7803344726562 }, { "name": "is_all_digits", "gain": 782.1705322265625 }, { "name": "in_wikipedia", "gain": 531.300048828125 }, { "name": "len", "gain": 489.9974365234375 }, { "name": "knnmp_log_max", "gain": 330.5997619628906 }, { "name": "n_clubs", "gain": 207.94166564941406 }, { "name": "n_letters", "gain": 196.677490234375 }, { "name": "origin__chinese", "gain": 170.69927978515625 }, { "name": "ends_digit", "gain": 165.2652587890625 }, { "name": "knnmp_count", "gain": 161.44723510742188 }, { "name": "n_digits", "gain": 159.4886016845703 }, { "name": "name_age_days", "gain": 156.08135986328125 }, { "name": "n_lower", "gain": 148.55072021484375 }, { "name": "n_special", "gain": 141.7484130859375 }, { "name": "is_palindrome", "gain": 139.62179565429688 }, { "name": "n_unique_chars", "gain": 124.55799102783203 }, { "name": "eth_stable_mcap", "gain": 124.44175720214844 }, { "name": "trademark_conflict", "gain": 116.59305572509766 }, { "name": "eth_tvl_usd", "gain": 107.47349548339844 }, { "name": "has_unicode", "gain": 106.94119262695312 }, { "name": "fame_score", "gain": 103.080078125 }, { "name": "knnbg_log_max", "gain": 102.78841400146484 }, { "name": "club__prepunk_full_rankings", "gain": 97.14898681640625 }, { "name": "is_ascii", "gain": 96.375 }, { "name": "club__gamertags", "gain": 94.76091766357422 }, { "name": "pca_045", "gain": 86.68685913085938 }, { "name": "pca_002", "gain": 85.17119598388672 }, { "name": "pca_004", "gain": 81.07220458984375 } ], "gain_split": { "mpnet_ft_knn": 4593.555641174316, "bge_knn": 324.30273818969727, "llm_name_kind": 197.1177396774292, "llm_cultural_origin": 421.92126178741455, "llm_scores": 103.080078125, "llm_description": 748.444676399231, "llm_total": 1470.5637559890747 }, "wandb_run": "https://wandb.ai/quantumly-aletheia-research/ens-appraiser/runs/m1mlf4h0" }