hannesthu

/

make_scale_by_max

like 0

• "__builtin__.set",
• "transformer_lens.components.Attention",
• "transformer_lens.hook_points.HookPoint",
• "transformer_lens.HookedTransformer.HookedTransformer",
• "torch._utils._rebuild_parameter",
• "torch.FloatStorage",
• "transformer_lens.components.Embed",
• "transformer_lens.HookedTransformerConfig.HookedTransformerConfig",
• "numpy.core.multiarray.scalar",
• "transformer_lens.components.PosEmbed",
• "torch.float32",
• "numpy.dtype",
• "torch._utils._rebuild_tensor_v2",
• "torch.nn.modules.container.ModuleList",
• "torch.BoolStorage",
• "transformer_lens.components.Unembed",
• "torch.nn.functional.relu",
• "transformer_lens.components.TransformerBlock",
• "transformer_lens.components.MLP",
• "_codecs.encode",
• "torch.device",
• "torch.nn.modules.linear.Identity",
• "collections.OrderedDict"

How to fix it?