{ "architectures": [ "Qwen3_5MoeForConditionalGeneration" ], "torch_dtype": "bfloat16", "image_token_id": 248056, "model_name": "/root/.cache/kagglehub/models/jackjirong2/qwen3.6-35b-a3b/transformers/default/1", "model_type": "qwen3_5_moe", "pad_token_id": 248044, "text_config": { "attention_bias": false, "attention_dropout": 0.0, "attn_output_gate": true, "bos_token_id": 248044, "torch_dtype": "bfloat16", "eos_token_id": 248044, "full_attention_interval": 4, "head_dim": 256, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "layer_types": [ "linear_attention", "linear_attention", "linear_attention", "full_attention", "linear_attention", "linear_attention", "linear_attention", "full_attention", "linear_attention", "linear_attention", "linear_attention", "full_attention", "linear_attention", "linear_attention", "linear_attention", "full_attention", "linear_attention", "linear_attention", "linear_attention", "full_attention", "linear_attention", "linear_attention", "linear_attention", "full_attention", "linear_attention", "linear_attention", "linear_attention", "full_attention", "linear_attention", "linear_attention", "linear_attention", "full_attention", "linear_attention", "linear_attention", "linear_attention", "full_attention", "linear_attention", "linear_attention", "linear_attention", "full_attention" ], "linear_conv_kernel_dim": 4, "linear_key_head_dim": 128, "linear_num_key_heads": 16, "linear_num_value_heads": 32, "linear_value_head_dim": 128, "mamba_ssm_dtype": "float32", "max_position_embeddings": 262144, "model_type": "qwen3_5_moe_text", "moe_intermediate_size": 512, "mtp_num_hidden_layers": 0, "mtp_use_dedicated_embeddings": false, "num_attention_heads": 16, "num_experts": 256, "num_experts_per_tok": 8, "num_hidden_layers": 40, "num_key_value_heads": 2, "output_router_logits": false, "pad_token_id": null, "partial_rotary_factor": 0.25, "rms_norm_eps": 1e-06, "rope_parameters": { "mrope_interleaved": true, "mrope_section": [ 11, 11, 10 ], "partial_rotary_factor": 0.25, "rope_theta": 10000000, "type": "default" }, "router_aux_loss_coef": 0.001, "shared_expert_intermediate_size": 512, "tie_word_embeddings": false, "use_cache": true, "vocab_size": 248320, "num_nextn_predict_layers": 0 }, "tie_word_embeddings": false, "unsloth_version": "2026.4.8", "video_token_id": 248057, "vision_end_token_id": 248054, "vision_start_token_id": 248053, "tool_parser_type": "qwen3_coder", "mtp_num_hidden_layers": 0, "num_nextn_predict_layers": 0, "quantization": { "group_size": 32, "bits": 4, "mode": "mxfp4", "language_model.model.layers.0.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.0.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.1.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.1.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.2.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.2.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.3.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.3.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.4.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.4.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.5.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.5.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.6.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.6.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.7.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.7.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.8.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.8.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.9.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.9.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.10.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.10.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.11.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.11.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.12.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.12.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.13.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.13.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.14.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.14.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.15.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.15.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.16.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.16.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.17.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.17.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.18.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.18.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.19.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.19.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.20.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.20.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.21.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.21.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.22.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.22.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.23.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.23.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.24.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.24.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.25.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.25.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.26.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.26.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.27.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.27.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.28.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.28.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.29.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.29.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.30.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.30.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.31.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.31.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.32.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.32.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.33.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.33.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.34.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.34.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.35.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.35.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.36.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.36.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.37.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.37.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.38.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.38.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.39.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.39.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" } }, "quantization_config": { "group_size": 32, "bits": 4, "mode": "mxfp4", "language_model.model.layers.0.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.0.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.1.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.1.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.2.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.2.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.3.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.3.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.4.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.4.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.5.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.5.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.6.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.6.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.7.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.7.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.8.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.8.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.9.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.9.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.10.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.10.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.11.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.11.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.12.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.12.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.13.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.13.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.14.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.14.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.15.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.15.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.16.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.16.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.17.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.17.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.18.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.18.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.19.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.19.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.20.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.20.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.21.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.21.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.22.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.22.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.23.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.23.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.24.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.24.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.25.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.25.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.26.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.26.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.27.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.27.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.28.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.28.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.29.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.29.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.30.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.30.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.31.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.31.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.32.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.32.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.33.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.33.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.34.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.34.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.35.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.35.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.36.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.36.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.37.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.37.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.38.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.38.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.39.mlp.gate": { "group_size": 64, "bits": 8, "mode": "affine" }, "language_model.model.layers.39.mlp.shared_expert_gate": { "group_size": 64, "bits": 8, "mode": "affine" } }, "abliteration": { "method": "residual_direction_weight_orthogonalization", "strength": 2.0, "targets": "attention,shared_down,switch_down", "layers": "", "direction_scope": "global", "preserve_column_norm": true, "edited_modules": [ { "name": "language_model.model.layers.39.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 39, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.39.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 39, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.39.self_attn.o_proj", "kind": "attention", "layer": 39, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.38.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 38, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.38.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 38, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.38.linear_attn.out_proj", "kind": "attention", "layer": 38, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.37.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 37, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.37.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 37, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.37.linear_attn.out_proj", "kind": "attention", "layer": 37, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.36.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 36, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.36.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 36, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.36.linear_attn.out_proj", "kind": "attention", "layer": 36, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.35.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 35, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.35.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 35, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.35.self_attn.o_proj", "kind": "attention", "layer": 35, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.34.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 34, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.34.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 34, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.34.linear_attn.out_proj", "kind": "attention", "layer": 34, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.33.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 33, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.33.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 33, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.33.linear_attn.out_proj", "kind": "attention", "layer": 33, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.32.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 32, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.32.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 32, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.32.linear_attn.out_proj", "kind": "attention", "layer": 32, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.31.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 31, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.31.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 31, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.31.self_attn.o_proj", "kind": "attention", "layer": 31, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.30.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 30, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.30.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 30, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.30.linear_attn.out_proj", "kind": "attention", "layer": 30, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.29.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 29, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.29.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 29, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.29.linear_attn.out_proj", "kind": "attention", "layer": 29, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.28.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 28, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.28.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 28, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.28.linear_attn.out_proj", "kind": "attention", "layer": 28, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.27.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 27, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.27.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 27, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.27.self_attn.o_proj", "kind": "attention", "layer": 27, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.26.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 26, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.26.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 26, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.26.linear_attn.out_proj", "kind": "attention", "layer": 26, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.25.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 25, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.25.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 25, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.25.linear_attn.out_proj", "kind": "attention", "layer": 25, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.24.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 24, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.24.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 24, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.24.linear_attn.out_proj", "kind": "attention", "layer": 24, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.23.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 23, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.23.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 23, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.23.self_attn.o_proj", "kind": "attention", "layer": 23, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.22.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 22, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.22.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 22, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.22.linear_attn.out_proj", "kind": "attention", "layer": 22, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.21.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 21, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.21.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 21, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.21.linear_attn.out_proj", "kind": "attention", "layer": 21, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.20.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 20, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.20.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 20, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.20.linear_attn.out_proj", "kind": "attention", "layer": 20, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.19.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 19, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.19.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 19, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.19.self_attn.o_proj", "kind": "attention", "layer": 19, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.18.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 18, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.18.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 18, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.18.linear_attn.out_proj", "kind": "attention", "layer": 18, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.17.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 17, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.17.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 17, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.17.linear_attn.out_proj", "kind": "attention", "layer": 17, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.16.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 16, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.16.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 16, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.16.linear_attn.out_proj", "kind": "attention", "layer": 16, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.15.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 15, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.15.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 15, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.15.self_attn.o_proj", "kind": "attention", "layer": 15, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.14.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 14, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.14.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 14, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.14.linear_attn.out_proj", "kind": "attention", "layer": 14, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.13.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 13, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.13.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 13, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.13.linear_attn.out_proj", "kind": "attention", "layer": 13, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.12.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 12, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.12.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 12, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.12.linear_attn.out_proj", "kind": "attention", "layer": 12, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.11.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 11, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.11.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 11, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.11.self_attn.o_proj", "kind": "attention", "layer": 11, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.10.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 10, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.10.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 10, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.10.linear_attn.out_proj", "kind": "attention", "layer": 10, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.9.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 9, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.9.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 9, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.9.linear_attn.out_proj", "kind": "attention", "layer": 9, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.8.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 8, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.8.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 8, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.8.linear_attn.out_proj", "kind": "attention", "layer": 8, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.7.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 7, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.7.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 7, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.7.self_attn.o_proj", "kind": "attention", "layer": 7, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.6.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 6, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.6.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 6, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.6.linear_attn.out_proj", "kind": "attention", "layer": 6, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.5.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 5, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.5.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 5, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.5.linear_attn.out_proj", "kind": "attention", "layer": 5, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.4.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 4, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.4.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 4, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.4.linear_attn.out_proj", "kind": "attention", "layer": 4, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.3.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 3, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.3.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 3, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.3.self_attn.o_proj", "kind": "attention", "layer": 3, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.2.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 2, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.2.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 2, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.2.linear_attn.out_proj", "kind": "attention", "layer": 2, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.1.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 1, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.1.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 1, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.1.linear_attn.out_proj", "kind": "attention", "layer": 1, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.0.mlp.shared_expert.down_proj", "kind": "shared_down", "layer": 0, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.0.mlp.switch_mlp.down_proj", "kind": "switch_down", "layer": 0, "axis": "output", "bits": 4, "mode": "mxfp4" }, { "name": "language_model.model.layers.0.linear_attn.out_proj", "kind": "attention", "layer": 0, "axis": "output", "bits": 4, "mode": "mxfp4" } ], "edited_module_count": 120, "direction_description": "completion-projected refusal direction; source direction file is not required for inference" }, "vision_config": { "deepstack_visual_indexes": [], "depth": 27, "dtype": "bfloat16", "hidden_act": "gelu_pytorch_tanh", "hidden_size": 1152, "in_channels": 3, "initializer_range": 0.02, "intermediate_size": 4304, "model_type": "qwen3_5_moe", "num_heads": 16, "num_position_embeddings": 2304, "out_hidden_size": 2048, "patch_size": 16, "spatial_merge_size": 2, "temporal_patch_size": 2 }, "publisher": "shiftedx" }