Spaces:
Paused
Paused
Commit History
-
auto: sync run_qwen_cross_profile_generation_synthesis.py
Search9a63f8a
-
auto: sync run_qwen_phase3_evidence_synthesis.py
Search2f72321 -
auto: sync run_qwen_phase3_layer_edge_audit.py
Searchc8d34d8 -
auto: sync run_qwen_phase3_evidence_synthesis.py
Search1bdab81 -
auto: sync run_qwen_injection_layer_ablation.py
Search136d537
-
auto: sync run_qwen_phase_probe.py
Search20ccbfa
-
auto: sync run_qwen_anchor_geometry_probe.py
Search4cbc2ee
-
auto: sync run_qwen_anchor_geometry_probe.py
Searchd257dbc
-
auto: sync run_qwen_anchor_concept_direction_map.py
Search02e003c -
auto: sync run_qwen_contradiction_diagnostic.py
Search076b659 -
auto: sync run_qwen_expanded_diagnostic.py
Search5804be7 -
auto: sync run_qwen_geometry_generation_calibration.py
Search0c967b8 -
auto: sync run_qwen_anchor_geometry_profile_sweep.py
Search607c4cf -
auto: sync run_qwen_anchor_layer_profile_map.py
Search8769d1f -
auto: sync run_qwen_anchor_carryover_probe.py
Search95d54ca -
auto: sync run_qwen_contradiction_diagnostic.py
Searchaaca540 -
auto: sync run_qwen_anchor_geometry_profile_sweep.py
Search3eb98bd -
auto: sync run_qwen_cross_profile_probe.py
Searchb6046d4 -
auto: sync run_qwen_future_influence_probe.py
Search2d32784
-
auto: sync run_qwen_geometry_generation_calibration.py
Search102df3b -
auto: sync run_qwen_anchor_layer_profile_map.py
Searchf65ece4
-
auto: sync run_qwen_motif_protocol.py
Search9f47153 -
auto: sync run_qwen_motif_protocol.py
Searchfa62175 -
auto: sync run_qwen_motif_protocol.py
Searchc4cddd0 -
auto: sync run_qwen_motif_protocol.py
Searchdb0d9bf
-
auto: sync run_external_moe_family_wikitext_compare.py
Search02aaa6d -
auto: sync run_external_moe_family_wikitext_compare.py
Searchb65859f -
auto: sync run_external_moe_family_wikitext_compare.py
Search1c73cc2
-
feat: add param-matched testformer mode
Search6536cc7 -
auto: sync run_testformer_wikitext_combo_remote.py
Searchf37be5a -
feat: add testformer wikitext combo runner
Search742c943
-
auto: sync run_qwen_200m_training.py
Searchd2569c7 -
auto: sync run_qwen_200m_training.py
Search2d6b208 -
auto: sync run_qwen_200m_training.py
Search60cc1a4
-
fix: increase epoch slider max to 500
Verified signature328f063 -
fix: replace app.py with file-based upload (no heredoc f-string issue)
Verified signaturea2b7259 -
feat: FOG stress ablation with new tasks (conditional, intersection, compose_add, multihop)
Verified signaturea4b762c -
sync: FOG micro+medium configs, stress tasks, fast pipeline
Search6ef010e -
fix: recognize uniform_small model type in run_experiment
Searcheb96110 -
fix: proper controls — loss masking, uniform baseline, multi-seed
Search35e5b8a -
fix: add tiny config (seq=32, d=128) for fast FOG ablation
Searchc3488d4 -
feat: add FOG ablation framework for motif-aware transformer experiments
Search019d725
-
fix: conditional GPU decorator + 2h CPU timeout for overnight runs
Searcha367cb2 -
fix: add --r1_ceiling CLI arg + metadata for single-domain H6 runs
Search3505a30 -
fix: raise subprocess timeout to 580s, set ATTN_IMPL=sdpa for VRAM savings
Search27e1d67 -
fix: configurable attn_implementation via ATTN_IMPL env var for VRAM savings
Searchceaea0a -
fix: increase ZeroGPU duration to 600s
Searchee87aa3 -
fix: add CLI metadata for 20domain scripts (underscore args)
Search75cd81e -
sync: update from main — H6 continuous bias, attention detection, auto-calibration, new scripts
Search78c132b
-
auto: sync run_qwen_anchor_carryover_probe.py
Searchcfaf8cc